Ask NYC Browse All Data Public Safety Transportation Health Buildings

Algorithmic Tools Compliance Report

"This is an annual report on algorithmic tools used by City agencies, collected under Local Law 35 of 2022 (LL 35). It includes descriptions of the tool's use and purpose, datasets used, and vendor involvement. The report is also published as a PDF on the OTI website: https://www.nyc.gov/content/oti/pages/reports.

The full text of LL 35 is available online: https://legistar.council.nyc.gov/LegislationDetail.aspx?ID=4265421&GUID=FBA29B34-9266-4B52-B438-A772D81B1CB5

An "algorithmic tool" is defined by the law as: "Any technology or computerized process that is derived from machine learning, artificial intelligence, predictive analytics, or other similar methods of data analysis, that is used to make or assist in making decisions about and implementing policies that materially impact the rights, liberties, benefits, safety or interests of the public, including their access to available city services and resources for which they may be eligible. Such term includes, but is not limited to tools that analyze datasets to generate risk scores, make predictions about behavior, or develop classifications or categories that determine what resources are allocated to particular groups or individuals, but does not include tools used for basic computerized processes, such as calculators, spellcheck tools, autocorrect functions, spreadsheets, electronic communications, or any tool that relates only to internal management affairs such as ordering office supplies or processing payments, and does not materially affect the rights, liberties, benefits, safety or interests of the public."

City Government Office of Technology and Innovation (OTI) Dataset jaw4-yuem 27 fields
DOWNLOAD CSV
Dataset fields
Showing 50 real records
Administration for Children's Services
Year: 2025 • Agency: Administration for Children's Services • Department: Division of Policy, Planning, and Analysis
Year
2025
Agency
Administration for Children's Services
Department
Division of Policy, Planning, and Analysis
Tool Name
Accelerated Safety Analysis Protocol (ASAP) Tool
Date First Use
2018/05
Updated
Yes
Purpose Type
Performance evaluation
Computation Type
Ranking
Autonomy
Informative
Frequency
Daily
Population Type
Individuals
Population Type Individual
NA
Population Type Other
NA
Website
Not specified
Tool Desc
Predictions of Severe Harm (identifying likelihood of substantiated allegations of physical or sex abuse within the next 24 months) are based on a machine learning methodology and are calculated for all children involved in active investigations early in the investigation (day 10). An investigation is assigned a numeric likelihood of this outcome based on the child in the case with the highest likelihood. The ACS Quality Assurance unit in the Division of Child Protection reviews about 3,000 active investigations annually, selecting those with the highest likelihood of severe harm.
Purpose Desc
The Quality Assurance (QA) unit in the Division of Child Protection at ACS has the capacity to review about 3,000 investigation cases out of about 50,000 investigations annually. ACS developed this predictive model to support the selection of cases for QA review. Open investigations involving children with the greatest likelihood to experience future severe harm – substantiated allegations of physical or sex abuse in the following 24 months – are selected for review. The tool does not support decisions about services or interventions for individuals or families involved with ACS, beyond the selection of the case for this additional QA review.

If the QA review team identifies gaps in routine, required documentation or practice, the team speaks with the field office conducting the investigation and follows up to make certain these gaps have been addressed. Scores are not shared with staff in the QA unit or the investigative unit. The model only supports the decision about which investigations are prioritized for review by the QA unit.
Updated Desc
Historically, the model made predictions only on investigation cases; however, over time, the proportion of child protection cases that are tracked to Collaborative Assessment, Response, Engagement & Support (CARES) has increased, and it has become necessary to identify the risk of severe harm for children in these cases as well. Updated model predictions include CARES cases.

The model was also retrained on new cases to tackle data drift resulting in part from policy changes and practices.

The model no longer collects identifying information; the only identifier is the case number.
Identifying Info
No identifying information is collected or disclosed
Data Training
ACS trained the model on ACS historic administrative data about closed investigations from 2012 to 2022. The training set included about 815,284 observations (80 percent of the historic data). The model was tested on closed investigations from January 2012 to November 2022 with 205,241 observations (20 percent).
Data Input
Predictions are based on administrative data about prior and current child welfare involvement including investigations triggered by a New York State Central Register call and time spent in foster care. Only ACS administrative data are used in the model.
Data Output
Rank ordered list of open investigation cases involving children with the highest likelihood to experience future severe harm, defined as substantiated allegations of physical or sex abuse in the following 24 months to be reviewed by a special QA Review Team.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Administration for Children's Services
Year: 2025 • Agency: Administration for Children's Services • Department: Division of Policy, Planning, and Analysis
Year
2025
Agency
Administration for Children's Services
Department
Division of Policy, Planning, and Analysis
Tool Name
Caseloads Projection
Date First Use
2024/07
Updated
Yes
Purpose Type
Resource allocation
Computation Type
Forecasting
Autonomy
Informative
Frequency
Quarterly
Population Type
Geographic space; Individuals
Population Type Individual
ACS staff
Population Type Other
NA
Website
Not specified
Tool Desc
The model estimates weekly caseloads by area for the next 12 months from the run date (day 14 of the prevention case).

The model is a time series model that forecasts caseloads citywide. The forecast analysis is performed at a more granular geography for the borough level. The model accounts for seasonality, trend, and variation in data to accurately project future caseloads.

Staffing decisions are made based on projected caseloads, historical attrition rate, and current team size.
Purpose Desc
Caseloads projections help identify how busy certain neighborhoods are likely to be in the next 12 months to optimize workforce distribution across different boroughs in the city. As social workers complete their training program, they are assigned to field offices. The right number of Child Protective Service Social Workers needs to be assigned to each borough based on the workload.
Updated Desc
ACS retrained the model on more recent ACS historic caseloads data from July 2021 to December 2024. The model was tested on caseloads from January 2025 to November 2025.
Identifying Info
No identifying information is collected or disclosed
Data Training
ACS trained the model was trained on ACS historic caseloads from July 2021 to December 2024 and tested on caseloads from January 2025 to November 2025 to forecast caseloads over the next 12-month period.
Data Input
Predictions are based on administrative data on investigations and Family Service Units’ involvement is extracted from CONNECTIONS (a child welfare computer system that provides for the documentation of information about families and children in New York State). A snapshot of all open investigations was taken at the start of the week (every Monday). This value is treated as the average caseload for the week. Only ACS administrative data are used in the model.
Data Output
Forecast of number of open investigations involving children by borough.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Administration for Children's Services
Year: 2025 • Agency: Administration for Children's Services • Department: Division of Policy, Planning, and Analysis
Year
2025
Agency
Administration for Children's Services
Department
Division of Policy, Planning, and Analysis
Tool Name
Family Team Conference Model (FTC)
Date First Use
2025/02
Updated
Tool was created in CY2025
Purpose Type
Resource allocation
Computation Type
Ranking
Autonomy
Supervised
Frequency
Monthly
Population Type
Individuals
Population Type Individual
Families ACS serves
Population Type Other
NA
Website
Not specified
Tool Desc
The Family Team Conference (FTC) is a structured planning conference facilitated by ACS with families. The Prevention Reassessment Planning FTC is held with families that have been engaged in prevention services for approximately 100 days. The purpose of this conference is to assess whether the current services are effectively meeting the family’s needs, determine whether additional support or intervention is necessary, and strengthen collaboration between ACS, the families they serve, and provider agencies to promote better outcomes.

The FTC model was developed for the Division of Prevention Services to support timely, data-informed decision-making for families engaged in child welfare prevention services. The model produces a risk score for each child approximately 100 days after the start of a new prevention service - well in advance of the FTC-Reassessment Planning Conference. This prediction estimates the likelihood that the child will be the subject of an indicated New York State Central Register investigation within the next 24 months from the observation date.
Purpose Desc
The Division of Prevention Services (DPS) uses these predictions to help identify and prioritize active prevention cases that may benefit from enhanced attention during the FTC Reassessment Planning Conference. By surfacing families with elevated predicted risk, the model enables senior staff in DPS to assign ACS Child and Family Specialists (CFS) to these conferences, where they may tailor discussions about targeted service adjustments or strengthened supports.

The FTC model supports the core purposes of the FTC Reassessment Planning Conference, which include:
1. Evaluating the Family’s Progress: Highlight families whose risk trajectory suggests the need for closer examination of progress over the previous six months.
2. Reassessing Risk and Safety: Provide data-driven insight into factors associated with elevated likelihood of future indicated allegations.
3. Adjusting Service Plans: Inform decisions on modifying service plans to address emerging needs, service gaps, or barriers to engagement.
4. Supporting Family Voice in Decision-Making: Ensure families have the opportunity to participate meaningfully in discussions informed by clear, evidence-based information.
5. Strengthening Collaboration Across Supports: Equip CFS staff and provider agencies with consistent, actionable information to better coordinate planning and interventions.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
ACS trained the model on ACS historic administrative data about closed investigations from 2014 to 2021. The training set contains 138,712 cases (80 percent of the historic data) ending between Jan 2014 and December 2021. The test set consisted of 34,862 cases (20 percent) ending between Jan 2014 and December 2021. The data was split such that no family appears in both sets.
Data Input
Predictions are based on administrative data about prior and current child welfare involvement at the start of a case. This includes New York State Central Register (SCR) investigations and time spent in foster care. Only ACS administrative data are used in the model.
Data Output
A monthly rank ordered list of open prevention cases involving children whose families have the highest likelihood of being in and indicated SCR investigation within 24 months from date of observation.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Administration for Children's Services
Year: 2025 • Agency: Administration for Children's Services • Department: Division of Policy, Planning, and Analysis
Year
2025
Agency
Administration for Children's Services
Department
Division of Policy, Planning, and Analysis
Tool Name
Housing Prioritization
Date First Use
2023/04
Updated
No
Purpose Type
Resource allocation
Computation Type
Ranking
Autonomy
Informative
Frequency
Monthly
Population Type
Individuals
Population Type Individual
NA
Population Type Other
NA
Website
Not specified
Tool Desc
The model estimates a risk score for a child receiving prevention services whose family will apply and be eligible for a homeless shelter within 12 months from the start of service (day 14 of the prevention case).

The model helps predict the risk of application for homeless shelters among families receiving prevention services. With a limited number of vouchers available, the risk model helps ACS prioritize housing assistance for those families at greatest risk of becoming homeless.

The service provider meets with the family to conduct a qualitative assessment of the family’s housing needs and vouchers are offered based on their findings. This is one of multiple ways in which ACS can identify families potentially eligible for housing assistance.
Purpose Desc
The city has allocated 100 housing vouchers to families receiving ACS prevention services. The shelter application model identifies the likelihood of a family in prevention services applying for homeless shelter within 12 months beyond the current prevention case. The model uses a machine learning methodology and is calculated for all children in a prevention case. ACS prevention services uses the results as one of several ways of identifying possible families that service providers can assist in applying for shelter.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
ACS trained the model on ACS historic administrative data regarding preventive services started between 2014 and 2020. The training set contains 140,242 observations (80 percent of the historical data) between January 2014 and December 2020. The test set consisted of 34,508 observations (20 percent) between January 2014 and December 2020. The data is split such that no family appears in both sets.
Data Input
Predictions are based on administrative data about prior and current child welfare involvement at the start of a case. This includes New York State Central Registry investigations and time spent in foster care. Only ACS administrative data are used in the model.
Data Output
Rank ordered list of open prevention cases involving children whose families have the highest likelihood of applying for a homeless shelter within 12 months of starting a prevention service.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Administration for Children's Services
Year: 2025 • Agency: Administration for Children's Services • Department: Division of Policy, Planning, and Analysis
Year
2025
Agency
Administration for Children's Services
Department
Division of Policy, Planning, and Analysis
Tool Name
Prevention Score Card
Date First Use
2021/09
Updated
Yes
Purpose Type
Performance evaluation
Computation Type
Ranking
Autonomy
Monitored
Frequency
Yearly
Population Type
Individuals
Population Type Individual
NA
Population Type Other
NA
Website
Not specified
Tool Desc
Predictions of Repeat Maltreatment (identifying the likelihood of being involved in a future indicated investigation within the next 24 months at the start of service) are based on a machine learning methodology and are calculated for all children receiving prevention services from ACS prevention service providers.
Purpose Desc
The Repeat Maltreatment model is used to make predictions on day 10 from the start of the prevention case to assess the risk of the family at the beginning of the service. A prevention case is assigned a numeric likelihood of an indicated investigation based on a New York State Central Register (SCR) within 24 months from the start of a prevention service.

The prevention providers are assessed for their performance based on the service needs/risk levels of the families they’ve served during the previous fiscal year.

The programs were sorted and ranked based on their average risk, and then divided into four quartiles by rank order: the top 25 percent of programs are classified as the Very High-Risk Cohort, the next 25 percent of programs as the High-Risk Cohort, the next 25 percent as the Medium-Risk Cohort, and the lowest 25 percent as the Low-Risk Cohort. Assignment to a cohort is not a way of performance assessment of the program but to group prevention service providers for fair comparisons based on the risk level of families they served.
Updated Desc
ACS retrained the model on recent ACS historic administrative data about closed investigations from July 2010 to June 2019. The training set included about 338,467 observations (80 percent of historic data). The model was tested on about 84,494 observations (20 percent). The tool was also updated such that identifying information is not included in the training, input, or output data, only a case number.
Identifying Info
No identifying information is collected or disclosed
Data Training
ACS trained the model on recent ACS historic administrative data about closed investigations from July 2010 to June 2019. The training set included about 338,467 observations (80 percent of historic data). The model was tested on about 84,494 observations (20 percent).
Data Input
Predictions are based on administrative data about prior and current child welfare involvement at the start of a case. This includes SCR investigations and time spent in foster care. Only ACS administrative data are used in the model.
Data Output
The model is used for generating a scorecard of prevention service providers by categorizing prevention programs based on the average risk profile of the cases they served during the assessment year. These groupings of program cohorts provide context for understanding the scorecard, as it allows for performance comparison of programs that accepted and served families with similar risk profiles.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Administration for Children's Services
Year: 2025 • Agency: Administration for Children's Services • Department: Division of Policy, Planning, and Analysis
Year
2025
Agency
Administration for Children's Services
Department
Division of Policy, Planning, and Analysis
Tool Name
Service Utilization Model
Date First Use
2025/01
Updated
Tool was created in CY2025
Purpose Type
Resource allocation
Computation Type
Forecasting
Autonomy
Supervised
Frequency
Biweekly
Population Type
Individuals
Population Type Individual
Families ACS serves
Population Type Other
NA
Website
Not specified
Tool Desc
Predictions of Repeat Maltreatment (identifying the likelihood of being involved in a future indicated investigation within the following 24 months towards the end of service) are based on a machine learning model and are calculated for all children receiving prevention services from ACS providers. A prevention case is assigned a numeric likelihood of an indicated investigation by the New York State Central Register (SCR) within 24 months of the end of a prevention service.
Purpose Desc
The Repeat Maltreatment model is a low-risk Service Utilization Model and is used by Department of Preventive Services (DPS) managers to identify the low-risk cases in prevention and to understand the length of service of these cases. DPS utilizes these reports to understand which programs have cases with a high length of service and are low risk. This report and the analysis are utilized internally for assessment purposes to understand program utilization and capacity.

The information from this report is not shared with providers.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
ACS trained the model on ACS historic administrative data about closed investigations from January 2014 to December 2021. The training set included about 138,712 observations (80 percent of the historic data). The model was tested on the remaining 34,862 observations (20 percent).
Data Input
Predictions are based on administrative data about prior and current child welfare involvement at the end of a prevention service case. It includes SCR investigations and time spent in foster care. Only ACS administrative data are used in the model.
Data Output
Rank ordered list of open prevention cases involving children with the lowest likelihood to experience repeat maltreatment, defined as substantiated allegations of abuse or neglect in the following 24 months to be reviewed by Preventive Services managers to identify the low-risk cases and to understand the length of service of these cases.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Campaign Finance Board
Year: 2025 • Agency: Campaign Finance Board • Department: Public Affairs
Year
2025
Agency
Campaign Finance Board
Department
Public Affairs
Tool Name
Adobe Creative Suite
Date First Use
2024/02
Updated
Yes
Purpose Type
Information presentation
Computation Type
Data transformation
Autonomy
Informative
Frequency
Project-based throughout the year
Population Type
Individuals
Population Type Individual
People engaging with NYC Votes creative ad campaigns and voter engagement and education materials across digital and print
Population Type Other
NA
Website
Not specified
Tool Desc
The CFB uses Adobe Creative Suite to edit images and uses some features that use generative AI to allow designers to edit images with reduced manual work.
Purpose Desc
The CFB uses Adobe Creative Suite to facilitate efficient graphic design that appear in some content, including NYC Votes creative ad campaigns and for voter engagement and education materials across digital and print. For example, Adobe Photoshop is used to remove figures in background photos who did not provide consent for usage. It is used to modify background elements that would be distracting (e.g., external telephone wires, shadows). As another example, the "Generative Recolor" feature in Adobe Illustrator allows designers to generate color palettes and apply those new colorways to patterns and artwork. Previously, designers needed to do this manually by individually clicking each element and applying a new color one at a time; "Generative Recolor" does this more efficiently.
Updated Desc
Improvements to photo generation/modification, vector graphic creation, design efficiency.
Identifying Info
No identifying information is collected or disclosed
Data Training
According to Adobe’s website, “generative AI Image models were trained on licensed content, such as Adobe Stock, and public domain content where copyright has expired.”
Data Input
Images.
Data Output
Visual content; primarily voter engagement and education materials across digital and print.
Vendor Name
Adobe
Vendor Type
Procurement/Paid pilot
Vendor Desc
This is a platform service.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Campaign Finance Board
Year: 2025 • Agency: Campaign Finance Board • Department: Public Affairs
Year
2025
Agency
Campaign Finance Board
Department
Public Affairs
Tool Name
Zoom
Date First Use
2020/07
Updated
No
Purpose Type
Information presentation
Computation Type
Data transformation
Autonomy
Fully autonomous
Frequency
Regularly/monthly
Population Type
Individuals
Population Type Individual
People who attend CFB public meetings
Population Type Other
NA
Website
Not specified
Tool Desc
Zoom is a virtual meeting platform; Zoom has an auto closed-caption function that uses AI.
Purpose Desc
The CFB uses Zoom for public meetings, including Board meetings, Voter Assistance Advisory Committee (VAAC) meetings, and All Partners meetings. The “automated captions” option is enabled on the agency’s Zoom account; Zoom-generated captions are available to attendees, who can choose to turn on this feature during the meeting.

(Note: For VAAC meetings and All Partner meetings, the agency also provides human-typed Communication Access Realtime Translation (CART) services as an option to all attendees. For All Partners meeting attendees, the CART transcription is made available after the call.)
Updated Desc
NA
Identifying Info
Input data; Output data
Data Training
Training data is proprietary to Zoom.
Data Input
Speech at the CFB’s public meetings.
Data Output
Text in a transcript and captions.
Vendor Name
Zoom
Vendor Type
Procurement/Paid pilot
Vendor Desc
This is a platform service.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of City Planning
Year: 2025 • Agency: Department of City Planning • Department: Community Planning and Civic Engagement
Year
2025
Agency
Department of City Planning
Department
Community Planning and Civic Engagement
Tool Name
Voice to Vision
Date First Use
2025/03
Updated
Tool was created in CY2025
Purpose Type
Information presentation
Computation Type
Data transformation
Autonomy
Informative
Frequency
Voice to Vision is viewed 10+ times per day
Population Type
Individuals; Geographic space; Property; Group, organization, or business
Population Type Individual
Voice to Vision shares individual comments shared during community planning
Population Type Other
NA
Website
https://dl.acm.org/doi/10.1145/3715070.3757228
Tool Desc
Voice to Vision is a repository of community input shared during neighborhood planning processes. Public comments were tagged and summarized by large language models and visualized online. This allows community members to see how their input shaped planning and policy decisions. As of December 2025, it has been used for the Jamaica Neighborhood Plan (https://v2v-jamaicaplan.ccc-mit.org/) and the White Plains Road Neighborhood Plan (https://v2v-white-plains-road-plan.ccc-mit.org/).
Purpose Desc
Voice to Vision uses artificial intelligence to support several functions for tagging and summarizing comments made during neighborhood planning sessions with the general public.Locations mentioned in community comments are identified and geotagged using GPT-4.1-mini and GPT-5-mini. Community comments are also tagged based on common themes using GPT-4.1-mini and GPT-5-mini. Finally, summaries of common topic areas are drafted by Claude 3.5 Sonnet, which also provides citations back to the original comments. All tagging, summarization, and citation is reviewed by staff prior to publication.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
Training data is proprietary to the vendor.
Data Input
Comments from community engagement sessions related to the neighborhood plans.
Data Output
Potential locations mentioned in comments, themes and goals mentioned in comments, theme summaries with citations from comments.
Vendor Name
Massachusetts Institute of Technology (MIT): The Center for Constructive Communication
Vendor Type
No-cost engagement
Vendor Desc
Voice to Vision was developed as a research project by MIT’s Center for Constructive Communication in coordination with DCP.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Correction
Year: 2025 • Agency: Department of Correction • Department: Operations Research
Year
2025
Agency
Department of Correction
Department
Operations Research
Tool Name
NYC Jail Population Forecast
Date First Use
2025/01
Updated
Tool was created in CY2025
Purpose Type
Resource allocation
Computation Type
Forecasting
Autonomy
Informative
Frequency
Weekly / Monthly
Population Type
Group, organization, or business
Population Type Individual
NA
Population Type Other
NA
Website
https://council.nyc.gov/budget/wp-content/uploads/sites/54/2025/09/DOC-MOCJ-FY25-TC-July-Report-Population-Forecast.pdf
Tool Desc
This is a two-year population projection for the New York City jail system based on historic trends from the past 12 years.
Purpose Desc
An older forecast model was created prior to this for internal informational and research purposes. However, in 2024, Terms and Conditions were added to DOC's budget requiring a more formal tool be developed and published for the jail system. The first of these reports was released in January 2025, and updates have been released every 6 months since. Internally, the Department has been using the tool for high-level planning and budget purposes.
Updated Desc
NA
Identifying Info
Training data; Input data
Data Training
Training data is from internal sources and includes historic aggregated jail admissions rates, NYPD daily arrest rates, daily average population counts, and information for incarcerated individuals including admission dates, historic changes in incarceration status (e.g., City Sentenced, Pretrial Detainee, State Ready), and length of stay.
Data Input
Inputs for inference include the latest aggregated population levels, admission rates, arrest rates, and information for incarcerated individuals including admission dates, historic changes in incarceration status, and length of stay.
Data Output
The system's output is a daily forecast for the expected population and the uncertainty around the daily forecast, for each day over the next two years. The uncertainty is estimated by the errors collected in a series of rigorous back-testing of the model and is shown as daily "prediction intervals" (i.e., the expected ranges for population) for each day over the same two-year period.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Correction
Year: 2025 • Agency: Department of Correction • Department: Division of Programs and Community Partnerships
Year
2025
Agency
Department of Correction
Department
Division of Programs and Community Partnerships
Tool Name
Pocketalk
Date First Use
2025/01
Updated
Tool was created in CY2025
Purpose Type
Information presentation
Computation Type
Data transformation
Autonomy
Supervised
Frequency
Limited assessments (6 total)
Population Type
Individuals
Population Type Individual
DOC staff, people in custody, and visitors to DOC facilities
Population Type Other
NA
Website
https://pocketalk.com/
Tool Desc
Pocketalk is designed to support real-time, two-way conversation translation across more than 90 languages. The Department of Correction has implemented a small-scale, limited pilot to assess the efficacy of the devices in a real-life environment. If successful, Pocketalk will be piloted across select facilities to supplement (not replace) qualified human interpretation services. Data captured through Pocketalk's use will assist the Department in monitoring usage trends, operational effectiveness, and potential expansion opportunities to enhance language accessibility across the agency.
Purpose Desc
Language translation for people in custody and visitors.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
Training data is proprietary to the vendor.
Data Input
Audio recordings of conversations.
Data Output
Written and spoken translation of the conversation into the specified language.
Vendor Name
Pocketalk
Vendor Type
Off-the-shelf
Vendor Desc
Pocketalk is a two-way translation device. Pocketalk utilizes multiple engines to provide translations across 92+ languages, including localized dialects and slang.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Correction
Year: 2025 • Agency: Department of Correction • Department: Division of Programs and Community Partnerships
Year
2025
Agency
Department of Correction
Department
Division of Programs and Community Partnerships
Tool Name
Risk of Recidivism Tool
Date First Use
2025/07
Updated
Tool was created in CY2025
Purpose Type
Resource allocation
Computation Type
Classification
Autonomy
Informative
Frequency
Approximately weekly
Population Type
Individuals
Population Type Individual
Individuals who are newly admitted to DOC custody
Population Type Other
NA
Website
Not specified
Tool Desc
The tool is designed to identify individuals at high risk of recidivism so they can receive additional services.
Purpose Desc
The purpose of the tool is to identify newly admitted individuals at high risk of recidivism within 12 months of discharge to the community so they can be offered targeted services that support successful reintegration. This initiative is funded by a Department of Justice grant that supports the hiring of two Discharge Planners to deliver these services, including weekly one-on-one case management, evidence-based curricula, and connections to community resources. The program is grounded in research demonstrating that enhanced, targeted interventions for individuals at high risk of recidivism lead to improved outcomes.
Updated Desc
NA
Identifying Info
Training data; Input data; Output data
Data Training
Information about persons in custody discharged to the community in 2020 and 2021.
Data Input
Prior admissions, self-reported homelessness, self reported drug or alcohol misuse, whether the top charge is a misdemeanor, and age of first admission to DOC custody.
Data Output
The tool outputs a classification indicator of either "high" or "low" risk of recidivism. Program staff then provide additional support to those who are at high risk.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Environmental Protection
Year: 2025 • Agency: Department of Environmental Protection • Department: Bureau of Business Information Technology
Year
2025
Agency
Department of Environmental Protection
Department
Bureau of Business Information Technology
Tool Name
AI Cybersecurity Tool
Date First Use
2025/07
Updated
Tool was created in CY2025
Purpose Type
Risk management
Computation Type
Classification
Autonomy
Monitored
Frequency
Not specified
Population Type
Group, organization, or business
Population Type Individual
NA
Population Type Other
NA
Website
Not specified
Tool Desc
DEP uses AI to analyze cybersecurity data to identify potential threats quickly and coordinate an effective response.
Purpose Desc
The information produced by the tool is used to identify unusual or potentially cybersecurity-risk activity across our environment quickly. DEP uses the insights to prioritize investigations, streamline cybersecurity operations, and respond more effectively to potential cyber threats.
Updated Desc
NA
Identifying Info
Input data; Output data
Data Training
Training data is proprietary to the vendor.
Data Input
Computer and network logs, including DEP staff user IDs and computer information.
Data Output
Alerts, risk scores, unusual network patterns, potential cybersecurity issues, and recommended actions, including DEP staff user IDs and computer information.
Vendor Name
Not disclosable
Vendor Type
Procurement/Paid pilot
Vendor Desc
Not disclosable due to cyber concerns and to prevent exposing details that could assist malicious actors.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Environmental Protection
Year: 2025 • Agency: Department of Environmental Protection • Department: Bureau of Sustainability
Year
2025
Agency
Department of Environmental Protection
Department
Bureau of Sustainability
Tool Name
FloodVision
Date First Use
2025/09
Updated
Tool was created in CY2025
Purpose Type
Risk management
Computation Type
Scoring
Autonomy
Fully autonomous
Frequency
Not specified
Population Type
Geographic space; Property; Group, organization, or business
Population Type Individual
NA
Population Type Other
NA
Website
Not specified
Tool Desc
FloodVision detects flooding in various areas of New York City by analyzing Department of Transportation (DOT) road camera footage and provides information about the flood to DEP to help inform data-driven decisions.
Purpose Desc
The FloodVision data is one of several data points that helps corroborate flooding hot spots in priority areas.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
Images and video from DOT road cameras.
Data Input
Images and video from DOT road cameras.
Data Output
Whether flooding has been detected and the approximate depth of the flood.
Vendor Name
ClearObject
Vendor Type
Professional services
Vendor Desc
ClearObject implemented FloodVision using Google Cloud Platform's AI cloud offerings.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Environmental Protection
Year: 2025 • Agency: Department of Environmental Protection • Department: Bureau of Business Information Technology
Year
2025
Agency
Department of Environmental Protection
Department
Bureau of Business Information Technology
Tool Name
Google Gemini
Date First Use
2025/09
Updated
Tool was created in CY2025
Purpose Type
Information presentation
Computation Type
Data transformation
Autonomy
Supervised
Frequency
Once
Population Type
Individuals
Population Type Individual
App users
Population Type Other
NA
Website
Not specified
Tool Desc
DEP used Gemini to generate an image that was edited and used as the icon for the NYC Noise app.
Purpose Desc
DEP used Gemini to help in part of a process to create a mobile app icon.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
Training data is proprietary to the vendor.
Data Input
A text prompt.
Data Output
An image that was combined with other images to create the app icon.
Vendor Name
Google
Vendor Type
Off-the-shelf
Vendor Desc
A free version of Google Gemini was used to create this graphic.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Environmental Protection
Year: 2025 • Agency: Department of Environmental Protection • Department: Public Affairs
Year
2025
Agency
Department of Environmental Protection
Department
Public Affairs
Tool Name
Microsoft Copilot for M365
Date First Use
2025/07
Updated
Tool was created in CY2025
Purpose Type
Information presentation
Computation Type
Data transformation
Autonomy
Informative
Frequency
Multiple times per week
Population Type
Group, organization, or business; Individuals
Population Type Individual
NA
Population Type Other
NA
Website
Not specified
Tool Desc
Microsoft Copilot for M365 is a tool that helps by drafting social media content, drafting scripts for video content, and providing translations of DEP content into other languages. This use was part of a pilot of Microsoft Copilot for M365 within Public Affairs, which began in July 2025 and is ongoing.
Purpose Desc
To assist in writing and presentation of agency objectives and public information for use on social media channels.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
Training data is proprietary to Microsoft, with access to data in the DEP Microsoft tenant.
Data Input
Staff prompts Copilot using DEP factual information and a desired tone for the output.
Data Output
Appropriate content for social media or video scripts; translated materials.
Vendor Name
Microsoft
Vendor Type
No-cost engagement
Vendor Desc
Microsoft provided the tool and some staff training.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Environmental Protection
Year: 2025 • Agency: Department of Environmental Protection • Department: Bureau of Environmental Compliance
Year
2025
Agency
Department of Environmental Protection
Department
Bureau of Environmental Compliance
Tool Name
Mufflers Noise Violations
Date First Use
2025/01
Updated
Tool was created in CY2025
Purpose Type
Performance evaluation
Computation Type
Classification
Autonomy
Supervised
Frequency
Daily
Population Type
Individuals; Geographic space
Population Type Individual
Owner of vehicle exceeding muffler noise limits
Population Type Other
NA
Website
Not specified
Tool Desc
Noise monitoring cameras record loud noises along with their noise decibel levels. The recordings are analyzed automatically to determine whether the type of noise recorded is from a car muffler, honking, or other types of noise. The software also tries to use Optical Character Recognition to read the vehicle license plates.
Purpose Desc
The information received is used by DEP staff to help inform whether a noise summons should be issued and to whom. While the tool helps with the analysis, the staff is still reviewing the recordings manually to verify the information provided.
Updated Desc
NA
Identifying Info
Input data; Output data
Data Training
Training data is proprietary to the vendor.
Data Input
Video recordings from installed cameras on New York City roads.
Data Output
Vehicle plate number and noise type.
Vendor Name
Intelligent Instruments
Vendor Type
Professional services
Vendor Desc
Intelligent Instruments provides the cameras that are installed in New York City roadways. Their software platform is doing the analysis on the recordings.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control – Public Health Laboratory
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control – Public Health Laboratory
Tool Name
Bowtie2
Date First Use
2022/06
Updated
No
Purpose Type
Data management
Computation Type
Matching
Autonomy
Supervised
Frequency
1-2 times per week
Population Type
Individuals; Biological sample
Population Type Individual
Patients who tested positive for the targeted pathogens and whose samples were sequenced
Population Type Other
NA
Website
https://github.com/BenLangmead/bowtie2
Tool Desc
Bowtie2 aligns sequencing data to a reference sequence using Burrows-Wheeler transformations. It is geared to use with Illumina sequencing data.
Purpose Desc
Bowtie2 is an intermediate step in the workflow to analyze COVID-19 variants in wastewater.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
N/A
Data Input
Laboratory-generated sequencing reads and an indexed organism reference genome.
Data Output
Genomic alignments indicating sequence variation relative to reference, supporting variant analysis and surveillance comparisons.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control – Public Health Laboratory
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control – Public Health Laboratory
Tool Name
Burrows-Wheeler Aligner (BWA)
Date First Use
2017/07
Updated
No
Purpose Type
Data management
Computation Type
Matching
Autonomy
Supervised
Frequency
1-2 times per week
Population Type
Individuals; Biological sample
Population Type Individual
Patients who tested positive for the targeted pathogens and whose samples were sequenced
Population Type Other
NA
Website
https://github.com/lh3/bwa
Tool Desc
Burrows-Wheeler Aligner (BWA) aligns sequencing data to a reference sequence.
Purpose Desc
BWA aligns sequence data to reference using Burrows-Wheeler transformations. This tool is optimal for low-divergent genomic data and short read data, such as Illumina sequence data. This tool is used to predict the order in which the fragments generated by sequencers are pieced together to form a complete genomic sequence data. This tool is used for Legionella and PulseNet sequencing analyses.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
N/A
Data Input
Laboratory-generated sequencing reads and an organism-specific reference genome selected for public health investigation.
Data Output
Read-to-reference alignments showing genomic similarity and variation, used for variant detection and comparison.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control – Bureau of Investigations
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control – Bureau of Investigations
Tool Name
ChoiceMaker (CM)
Date First Use
2003/06
Updated
No
Purpose Type
Data management
Computation Type
Matching
Autonomy
Supervised
Frequency
Daily
Population Type
Individuals
Population Type Individual
Individuals of all ages with a record in the Citywide Immunization Registry
Population Type Other
NA
Website
https://www.choicemaker.com/
Tool Desc
ChoiceMaker (CM) is a record-matching tool that identifies duplicate records belonging to the same individual.
Purpose Desc
CM is used by the Bureau of Investigations (BOI) and Healthy Homes to identify duplicate immunization and lead records. The outputs produced by CM are used in ongoing manual and automated deduplication processes (record merging).
Updated Desc
NA
Identifying Info
Training data; Input data; Output data
Data Training
The training data was developed by the vendor and DOHMH staff using DOHMH data. It includes DOHMH staff decisions on the likelihood of a match between 2 records.
Data Input
CM uses demographic data (e.g., names, date of birth, address, identifiers) and health event data (e.g., date and type of event) from BOI’s Citywide Immunization Registry and Healthy Homes’ LeadQuest registry in its evaluation.
Data Output
The program outputs a series of record pairs and a match probability for each pair.
Vendor Name
HLN Consulting
Vendor Type
Professional services
Vendor Desc
A vendor was involved in the development of the program initially. CM is now available as an open-source program. The DOHMH implementation, which has been customized using the training data, is maintained by HLN Consulting.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control - Public Health Laboratory
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control - Public Health Laboratory
Tool Name
Genome Analysis Toolkit (GATK)
Date First Use
2017/10
Updated
No
Purpose Type
Data management
Computation Type
Classification
Autonomy
Supervised
Frequency
3-4 times per week
Population Type
Individuals; Biological sample
Population Type Individual
Patients who tested positive for the targeted pathogens and whose samples were sequenced
Population Type Other
NA
Website
https://gatk.broadinstitute.org/hc/en-us
Tool Desc
The Genome Analysis Toolkit (GATK) is a suite of tools for variant calling and filtering after sequence alignment. It uses naive Bayes to qualify aligned bases as sequence or erroneous data, which would be excluded from the final genomic sequence.
Purpose Desc
GATK is used to identify mutations and call upon differences from the reference, which is used to generate the predicted complete sequence.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
Uses known variant site resources curated by public consortia for recalibration; no local model training performed.
Data Input
Reference genome and aligned sequencing data generated through public health laboratory workflows.
Data Output
Identified genomic variants relative to reference sequence, used for mutation tracking, resistance analysis, and outbreak comparison.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control - Public Health Laboratory
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control - Public Health Laboratory
Tool Name
Guppy
Date First Use
2020/06
Updated
No
Purpose Type
Data management
Computation Type
Classification
Autonomy
Supervised
Frequency
A few times per month
Population Type
Individuals; Biological sample
Population Type Individual
Patients who tested positive for the targeted pathogens and whose samples were sequenced
Population Type Other
NA
Website
https://nanoporetech.com/software/other/guppy
Tool Desc
Guppy converts electric signals to predict a nucleotide and enables filtering of low-quality calls.
Purpose Desc
This is a tool designed specifically for Oxford Nanopore Technology data. This is a neural network based basecaller: a tool that determines nucleotide bases of a genetic material and converts electric signals into strings to represent genomic data. In addition to basecalling, the tool also performs filtering of low-quality reads (a stretch of sequenced genetic material). This is the initial step that converts electric signals to fragments of sequence data, which can then be used for COVID-19 sequencing analysis.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
Vendor-developed basecalling models are trained by Oxford Nanopore Technologies on proprietary nanopore signal datasets representing electrical signal-to-nucleotide mappings.
Data Input
Raw electrical signal data generated by Oxford Nanopore sequencing instruments from clinical or surveillance specimens.
Data Output
Basecalled nucleotide sequences with quality metrics representing specimen genetic content, used for downstream genomic surveillance and outbreak analyses.
Vendor Name
Oxford Nanopore Technologies
Vendor Type
Off-the-shelf
Vendor Desc
Developed and maintains the tool.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Division of Information Technology
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Division of Information Technology
Tool Name
HELIX
Date First Use
2025/11
Updated
Tool was created in CY2025
Purpose Type
Information presentation
Computation Type
Data transformation
Autonomy
Informative
Frequency
Daily
Population Type
Individuals
Population Type Individual
Agency staff
Population Type Other
NA
Website
Not specified
Tool Desc
HELIX is a generative tool that allows agency staff to ask general questions based on data available in the OpenAI GPT-4o model. In addition, HELIX can also be grounded on approved documents allowing staff to ask questions specific to those documents.
Purpose Desc
The tool is used to generate responses to prompts entered by staff,
requesting the generative AI tool to compose a text response to a text input. These outputs are potentially shared with the public.
Updated Desc
NA
Identifying Info
Input data; Output data
Data Training
Training data is proprietary to the vendor.
Data Input
Prompts provided by the users of the system.
Data Output
The output data for the tool is the response generated by the OpenAI large language model. In addition, any HELIX configurations grounded on approve documents will generate responses based on those documents.
Vendor Name
Microsoft
Vendor Type
Off-the-shelf
Vendor Desc
Helix was built using Microsoft Azure OpenAI.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control – Bureau of Immunization
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control – Bureau of Immunization
Tool Name
Immunization Calculation Engine (ICE)
Date First Use
1997/00
Updated
Yes
Purpose Type
Information presentation
Computation Type
Forecasting
Autonomy
Monitored
Frequency
Daily
Population Type
Individuals
Population Type Individual
Anyone who needs a vaccine
Population Type Other
NA
Website
https://cdsframework.atlassian.net/wiki/spaces/ICE/overview
Tool Desc
The Immunization Calculation Engine (ICE) is an immunization evaluation and forecasting system whose default immunization schedule supports all routine childhood, adolescent, and adult immunizations based on the recommendations of the Advisory Committee on Immunization Practices. ICE is free and open-source.
Purpose Desc
ICE is used by the Bureau of Immunization to evaluate a patient’s immunization history and generate appropriate immunization recommendations.
Updated Desc
Vaccine recommendations were updated in CY2025.
Identifying Info
Input data
Data Training
N/A
Data Input
ICE uses demographic data (e.g., date of birth) and vaccination data (e.g., immunization date, vaccine group and type) in the evaluation process. Data used are stored in the Citywide Immunization Registry.
Data Output
The program returns recommendations on whether a patient has completed a vaccine series or is due for vaccines.
Vendor Name
HLN Consulting
Vendor Type
Professional services
Vendor Desc
A vendor was involved in the development of the program and continues to be involved in program enhancements. ICE is also available as an open-source program. The DOHMH implementation is maintained by HLN Consulting.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Bureau of Communicable Disease
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Bureau of Communicable Disease
Tool Name
Improving Foodborne Disease Outbreak Detection by Incorporating Complaints Identified in Social Media Data
Date First Use
2016/11
Updated
Yes
Purpose Type
Triage
Computation Type
Scoring
Autonomy
Informative
Frequency
Daily
Population Type
Individuals
Population Type Individual
Public who dine at New York City restaurants and are Yelp users and New York City restaurants
Population Type Other
NA
Website
https://publichealth.cs.columbia.edu/
Tool Desc
Restaurant-associated foodborne disease outbreaks are often identified through complaints received via 311 non-emergency information system; however not all individuals report to 311. DOHMH in collaboration with Columbia University developed a text classifier program which monitors Yelp and Twitter data to identify complaints of foodborne illness which was supported by grants from the Alfred P. Sloan Foundation and the National Science Foundation. As of April 2023, the tool no longer uses data from Twitter (X) due to application programming interface (API) changes.
Purpose Desc
The model uses data from Yelp restaurant reviews and previously used Twitter data that was available on Twitter’s publicly available API. Twitter (X) removed free access to their publicly available API in April 2023, so these data are no longer included in our analyses. The classifiers assign a “sick score” to each Yelp review indicating the likelihood that the review pertains to foodborne illness. The sick score is based on whether the review contains key words indicative of foodborne illness (e.g. “vomit”); the Yelp classifier also incorporates if the review indicates that multiple people became sick and if the review indicates a time between eating at a restaurant and illness onset (incubation period) that is consistent with foodborne illness. Each review with a sick score greater than or equal to a threshold value are reviewed and annotated by DOHMH foodborne disease epidemiology and environmental health staff to determine if the review was actually reporting foodborne illness possibly associated with a New York City restaurant; if yes, Yelp messages are sent to Yelp reviewers, requesting that they contact DOHMH. Data from annotations are used to improve classifier performance. Foodborne disease complaints identified through Yelp are combined with foodborne disease complaints reported to 311 to improve efficiency of outbreak detection.
Updated Desc
Since January 31, 2025, DOHMH has been unable to receive data for this tool. Columbia University and DOHMH staff have been unable to resolve the issue without additional resources.
Identifying Info
Training data; Input data
Data Training
Training data was used in the development of the Yelp classifiers. The training data consisted of restaurant reviews obtained from Yelp by Columbia University; a subset of these data were joined with annotations provided by DOHMH staff. The annotations of restaurant reviews focused on the following: 1) if the review indicated foodborne illness; 2) if the incident occurred in the past 30 days; 3) if multiple people were sick; and 4) if the incubation period was consistent with foodborne illness. The training data is periodically updated (with annotations from DOHMH) to improve the classifiers.
Data Input
Yelp reviews of New York City restaurants are pulled from a privately available API provided by Yelp.
Data Output
The output data includes a “sick score” that the classifiers assign to each Yelp review indicating the likelihood that the review pertains to foodborne illness.
Vendor Name
Columbia University
Vendor Type
No-cost engagement
Vendor Desc
DOHMH staff, including Bureau of Communicable Disease, Office of Environmental Investigations, and Division of Informatics and Information Technology & Telecommunications, and Columbia University are involved in making decisions about the tool. Columbia University Department of Computer Science professors and doctoral students maintain the classifier. The project was previously funded by the Alfred P. Sloan Grant, for which The Fund for Public Health in New York provided administrative support and grant management to DOHMH. This support and management ended at the completion of the grant in 2021.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control - Public Health Laboratory
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control - Public Health Laboratory
Tool Name
IQTREE
Date First Use
2020/05
Updated
No
Purpose Type
Data management
Computation Type
Clustering
Autonomy
Supervised
Frequency
2-3 times per week
Population Type
Individuals; Biological sample
Population Type Individual
Patients who tested positive for the targeted pathogens and whose samples were sequenced
Population Type Other
NA
Website
https://github.com/iqtree/iqtree3
Tool Desc
IQTREE uses maximum-likelihood regression to create phylogenetic trees from genomes.
Purpose Desc
Produced phylogenetic trees are used to help rule in or out outbreaks of COVID-19 or other organisms.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
N/A
Data Input
Multiple sequence alignments derived from public health genomic datasets.
Data Output
Phylogenetic trees representing evolutionary relationships among isolates, used to assess transmission clusters and outbreak linkage.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control - Public Health Laboratory
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control - Public Health Laboratory
Tool Name
kSNP4
Date First Use
2022/03
Updated
No
Purpose Type
Data management
Computation Type
Clustering
Autonomy
Supervised
Frequency
2-3 times per week
Population Type
Individuals; Biological sample
Population Type Individual
Patients who tested positive for the targeted pathogens and whose samples were sequenced
Population Type Other
NA
Website
https://pubmed.ncbi.nlm.nih.gov/37948764/
Tool Desc
kSNP4 uses multiple algorithms (e.g., maximum-likelihood, parsimony, neighbor-joining) to infer phylogenetic trees from genomes.
Purpose Desc
Produced phylogenetic trees are used to help rule in or out outbreaks of bacteria.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
N/A
Data Input
Assembled genome sequences generated from laboratory sequencing data.
Data Output
Single-nucleotide polymorphisms-based relatedness results and phylogenetic trees summarizing genetic similarity among isolates, used for outbreak investigation.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control - Public Health Laboratory
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control - Public Health Laboratory
Tool Name
Minimap2
Date First Use
2020/05
Updated
No
Purpose Type
Data management
Computation Type
Matching
Autonomy
Supervised
Frequency
3-4 times per week
Population Type
Individuals; Biological sample
Population Type Individual
Patients who tested positive for the targeted pathogens and whose samples were sequenced
Population Type Other
NA
Website
https://github.com/lh3/minimap2
Tool Desc
Minimap2 aligns sequencing data to a reference sequence.
Purpose Desc
Minimap2 uses optimal chaining scores to align sequencing data to reference genomes. This tool is faster and more optimal for long read sequences, such as Oxford Nanopore Technologies data. This tool is used to predict the order in which the fragments generated by sequencers are pieced together to form a complete genomic sequence data. This tool is used for COVID-19 and monkeypox virus sequencing analyses.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
N/A
Data Input
Laboratory-generated sequencing reads and an appropriate reference genome.
Data Output
Sequence alignments identifying genomic similarities and differences, used for genomic characterization and surveillance.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control - Public Health Laboratory
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control - Public Health Laboratory
Tool Name
Multiple Alignment using Fast Fourier Transform (MAFFT)
Date First Use
2021/01
Updated
No
Purpose Type
Data management
Computation Type
Matching
Autonomy
Supervised
Frequency
3-4 times per week
Population Type
Individuals; Biological sample
Population Type Individual
Patients who tested positive for the targeted pathogens and whose samples were sequenced
Population Type Other
NA
Website
https://github.com/GSLBiotech/mafft
Tool Desc
Multiple Alignment using Fast Fourier Transform (MAFFT) aligns multiple sequencing data.
Purpose Desc
MAFFT includes several algorithmic methods, including guided tree, scoring matrices, and sequence alignment algorithms to realign multiple genomic sequencing data. The tool aligns sequences to help identify differences. This is used in all sequencing analysis prior to building a phylogenetic tree or distance tree.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
N/A
Data Input
Homologous nucleotide or protein sequences generated from laboratory sequencing or validated reference databases.
Data Output
Aligned sequences identifying conserved and variable genomic regions, used for phylogenetic and outbreak analyses.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control - Public Health Laboratory
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control - Public Health Laboratory
Tool Name
Pangolin
Date First Use
2021/07
Updated
Yes
Purpose Type
Data management
Computation Type
Clustering
Autonomy
Supervised
Frequency
3-4 times per week
Population Type
Individuals; Biological sample
Population Type Individual
Patients who tested positive for the targeted pathogens and whose samples were sequenced
Population Type Other
NA
Website
https://github.com/cov-lineages/pangolin
Tool Desc
Pangolin assigns lineage names to SARS-CoV-2.
Purpose Desc
Pangolin uses a combination of several methods, including random forest tree, classification methods, and maximum parsimony to assign lineage names to SARS-CoV-2 genomic sequences to bin sequences that are more likely to be similar. This is a tool that designates a name based on a nomenclature for COVID-19 sequence data.
Updated Desc
Updated to pangolin version 4.3.3.
Identifying Info
No identifying information is collected or disclosed
Data Training
Curated SARS-CoV-2 genome dataset with Pango lineage designations is maintained by the Pango Network and periodically updated.
Data Input
SARS-CoV-2 consensus genome sequences generated through public health sequencing workflows.
Data Output
Viral lineage assignments indicating genetic classification, used for variant tracking and epidemiological reporting.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control - Public Health Laboratory
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control - Public Health Laboratory
Tool Name
PHYLOViZ
Date First Use
2017/10
Updated
No
Purpose Type
Data management
Computation Type
Clustering
Autonomy
Supervised
Frequency
3-4 times per year or as required
Population Type
Individuals; Biological sample
Population Type Individual
Patients who tested positive for the targeted pathogens and whose samples were sequenced
Population Type Other
NA
Website
https://www.phyloviz.net/
Tool Desc
For representing the possible evolutionary relationships between strains, PHYLOViZ uses the goeBURST algorithm, a refinement of eBURST algorithm by Feil et al., and its expansion to generate a complete minimum spanning tree.
Purpose Desc
PHYLOViZ is used to generate the minimum spanning tree relationships.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
N/A
Data Input
Allelic profiles or phylogenetic outputs generated from sequencing-based typing analyses.
Data Output
Graphical relatedness networks identifying clustered isolates, used for outbreak visualization and surveillance reporting.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control - Public Health Laboratory
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control - Public Health Laboratory
Tool Name
PulseNet 2.0
Date First Use
2017/09
Updated
No
Purpose Type
Data management
Computation Type
Clustering
Autonomy
Supervised
Frequency
4-5 times per week
Population Type
Individuals; Biological sample
Population Type Individual
Patients with culture-positive foodborne bacterial isolate
Population Type Other
NA
Website
https://stacks.cdc.gov/view/cdc/138367
Tool Desc
PulseNet 2.0 is a suite of tools used to align and analyze bacterial genomes.
Purpose Desc
PulseNet 2.0 is used to:
1. Assemble the bacterial genome (since the sequencing process involves fragmenting the bacterial DNA and then amplifying it into millions of pieces);
2. Identify the genus, species, and serotype of the bacterial isolate;
3. Perform quality control checks to ensure the sequence meets certain quality standards;
4. Perform core and whole genome multi-locus sequence typing (a technique used to type bacteria based on their genetic code);
5. Perform cluster analysis for cases related to one another based upon case definitions recommended by the Centers for Disease Control and Prevention (CDC).
This information is then communicated to partners including foodborne epidemiologists at the Bureau of Communicable Disease, who investigate all reported cases of foodborne disease, with those investigations potentially resulting in restaurant inspections, closures, and food recalls.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
N/A
Data Input
Standardized sequencing data submitted by public health laboratories.
Data Output
Nationally comparable relatedness and clustering results used for multi-jurisdictional outbreak detection and surveillance coordination.
Vendor Name
Centers for Disease Control and Prevention, Booz Allen Hamilton
Vendor Type
Procurement/Paid pilot
Vendor Desc
Developed and maintains the tool.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control - Public Health Laboratory
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control - Public Health Laboratory
Tool Name
Spades
Date First Use
2017/10
Updated
No
Purpose Type
Data management
Computation Type
Matching
Autonomy
Supervised
Frequency
2-3 times per week
Population Type
Individuals; Biological sample
Population Type Individual
Patients who tested positive for the targeted pathogens and whose samples were sequenced
Population Type Other
NA
Website
https://github.com/ablab/spades
Tool Desc
Spades uses several algorithms to simplify genomic read data into de Brujin graphs and finds overlaps to assemble genomes.
Purpose Desc
Spades is an intermediate step in the workflows of bacterial analyses.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
N/A
Data Input
Raw sequencing reads produced from clinical, environmental, or foodborne pathogen samples.
Data Output
Draft genome assemblies reconstructing organism genomes, used for typing, resistance detection, and surveillance analyses.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Health and Mental Hygiene
Year: 2025 • Agency: Department of Health and Mental Hygiene • Department: Disease Control - Public Health Laboratory
Year
2025
Agency
Department of Health and Mental Hygiene
Department
Disease Control - Public Health Laboratory
Tool Name
Vsearch
Date First Use
2022/06
Updated
No
Purpose Type
Data management
Computation Type
Matching
Autonomy
Supervised
Frequency
As required per week/month
Population Type
Individuals; Biological sample
Population Type Individual
Patients who tested positive for the targeted pathogens and whose samples were sequenced
Population Type Other
NA
Website
https://github.com/torognes/vsearch
Tool Desc
Vsearch uses the Needleman-Wunsch algorithm to merge read pairs and align and dereplicate sequences to detect chimeric genomic sequences.
Purpose Desc
Vsearch is an intermediate step in the workflow to analyze COVID-19 variants in wastewater.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
N/A
Data Input
Nucleotide sequences from laboratory samples and, where applicable, curated reference databases.
Data Output
Similarity or clustering results identifying related sequences, used for organism identification or comparative genomic analysis.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Investigation
Year: 2025 • Agency: Department of Investigation • Department: Compliance
Year
2025
Agency
Department of Investigation
Department
Compliance
Tool Name
Facial Recognition Technology
Date First Use
2019/03
Updated
Yes
Purpose Type
Not specified
Computation Type
Matching
Autonomy
Informative
Frequency
Not specified
Population Type
Individuals
Population Type Individual
NA
Population Type Other
NA
Website
Not specified
Tool Desc
The tool analyzes an uploaded image or video and searches and compares it with lawfully possessed images to generate a pool of possible matches. If possible matches are identified, a trained DOI examiner visually analyzes and evaluates potential matches to assess reliability of a match consistent with agency policy and applicable laws. A match serves as an investigative lead for additional investigative steps and does not constitute a positive identification.
Purpose Desc
Facial recognition is a digital technology that DOI uses to analyze uploaded images or videos of people and objects obtained during an investigation by comparison with lawfully possessed images. Facial recognition generates possible matches of an object or individual from this analysis and comparison. The purpose of the tool is to assist DOI investigations of matters within its jurisdiction including fraud and other criminal activity.
Updated Desc
Routine vendor updates.
Identifying Info
Training data; Input data; Output data
Data Training
Vendor uses publicly available open-source media data.
Data Input
Images or videos of people and object obtained during an investigation.
Data Output
Lawfully possessed images of possible matches, if found.
Vendor Name
Not disclosable
Vendor Type
Off-the-shelf
Vendor Desc
The vendors provide ongoing technical assistance. Confidentiality agreements are in place with the vendors.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Social Services
Year: 2025 • Agency: Department of Social Services • Department: Human Resources Administration Homebase
Year
2025
Agency
Department of Social Services
Department
Human Resources Administration Homebase
Tool Name
Homebase Risk Assessment Questionnaire (RAQ)
Date First Use
2012/06
Updated
No
Purpose Type
Resource allocation
Computation Type
Scoring
Autonomy
Monitored
Frequency
Daily
Population Type
Individuals
Population Type Individual
Households seeking Homebase assistance
Population Type Other
NA
Website
Not specified
Tool Desc
Homebase applicants answer screening questions about their current housing situation, history of disruptive experiences, shelter history, and other domains. Each of the answers is assigned a number of points, and applicants that reach a certain point threshold are eligible for deeper Homebase services, such as financial assistance and case management. Workers are able to override a limited number of model decisions with permission of a supervisor.
Purpose Desc
The Homebase program was created to prevent households from entering the Department of Homeless Services (DHS) shelter system. Since New York City has a range of antipoverty programs and the number of households entering shelter is small compared to the pool of New Yorkers who have an eviction filing each year, the Agency had to ensure that the households who most needed additional homelessness prevention services were being enrolled in Homebase programs. Research showed that staff were not accurately able to predict who would or would not enter the DHS shelter system and that using a risk assessment would provide a better way to match resources to the families who would benefit the most.
Updated Desc
NA
Identifying Info
Training data; Input data; Output data
Data Training
The RAQ was developed based on analysis of data on Homebase enrollees from 2004 to 2008, conducted in conjunction with a team of academic researchers, to determine predictive factors for those entering shelter. It was updated in 2023 based on analysis led by DSS researchers of 2013-2016 Homebase data.
Data Input
Factors include, among others: personal characteristics such as age and pregnancy; educational attainment and employment status; housing issues such as eviction, discord, a move in the past year; past and recent experience of homelessness.
Data Output
The tool produces a score that is used to assess eligibility for full versus brief Homebase services.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Social Services
Year: 2025 • Agency: Department of Social Services • Department: Accountability Office
Year
2025
Agency
Department of Social Services
Department
Accountability Office
Tool Name
Maestro Automated Valuation System (MAVS)
Date First Use
2025/01
Updated
Tool was created in CY2025
Purpose Type
Triage
Computation Type
Scoring
Autonomy
Informative
Frequency
Regularly
Population Type
Individuals
Population Type Individual
Medicaid clients who file personal injury lawsuits
Population Type Other
NA
Website
Not specified
Tool Desc
Maestro Automated Valuation System (MAVS) uses machine learning to identify Medicaid encounters that are related to a given injury for which a Medicaid client received funds from a personal injury lawsuit. The Medicaid program is legally entitled to place a lien to recoup these medicaid funds.
Purpose Desc
MAVS uses machine learning to identify Medicaid encounters that are related to a given injury for which a Medicaid client received funds from a personal injury lawsuit. The Medicaid program is legally entitled to place a lien to recoup these Medicaid funds. Under current practice, a DSS staff member reviews MAVS output before sending any notification to the client or attorneys in the case.
Updated Desc
NA
Identifying Info
Training data; Input data; Output data
Data Training
The model was trained on MAESTRO data (the system storing the DSS medicaid lien data) and specifically on the determination decisions made manually by DSS staff members.
Data Input
Medicaid data warehouse encounter data and MAESTRO data on personal injury details.
Data Output
Each Medicaid encounter in the client's MAESTRO file receives a determination: "RELATED" or "UNRELATED."
Vendor Name
Gainwell/Health Management Systems, Inc. via contract with NYS Office of Medicaid Inspector General (OMIG)
Vendor Type
Professional services
Vendor Desc
DSS and NYS OMIG meet with Gainwell regularly to discuss MAVS and the underlying MAESTRO system.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Social Services
Year: 2025 • Agency: Department of Social Services • Department: Public Engagement Unit
Year
2025
Agency
Department of Social Services
Department
Public Engagement Unit
Tool Name
SmartVAN / TargetSmart
Date First Use
2019/11
Updated
No
Purpose Type
Data management
Computation Type
Matching
Autonomy
Informative
Frequency
Occasionally
Population Type
Individuals
Population Type Individual
New Yorkers potentially eligible for benefits
Population Type Other
NA
Website
Not specified
Tool Desc
The Mayor’s Public Engagement Unit (PEU) uses SmartVAN to manage outreach across a range of projects. SmartVAN provides functionality to create lists of potential clients to contact, collect personal information and survey responses from clients, and conduct outreach via phone banks and canvassing. SmartVAN also contains a frequently updated commercial dataset, provided by TargetSmart, of New York City residents and their demographic, contact, and other information. PEU uses this preloaded data to create outreach lists when data on existing clients or from partner agencies is unavailable or insufficient to meet the scope of the outreach project.
Purpose Desc
In 2025, PEU used TargetSmart data within SmartVAN on various outreach projects. For instance, PEU used it to pull lists of New Yorkers who recently turned 26 for outreach to offer help enrolling in health insurance.
Updated Desc
NA
Identifying Info
Training data; Input data; Output data
Data Training
Training data is part of vendor’s proprietary processes.
Data Input
Input data is part of vendor’s proprietary processes.
Data Output
The algorithmically-derived data that PEU accesses is the output of proprietary algorithmic processes developed and operated by TargetSmart. These algorithmic processes include matching multiple input datasets to determine residency, contact information, and demographics on New York City residents. SmartVAN also includes a number of algorithmically determined likelihood scores, including scores for the likelihood that a household contains children under 18, etc.
Vendor Name
EveryAction and TargetSmart
Vendor Type
Procurement/Paid pilot, Professional services
Vendor Desc
EveryAction and TargetSmart jointly provide the SmartVAN product. EveryAction is the software provider. TargetSmart is the data provider. TargetSmart is the entity who applies algorithmic techniques. EveryAction provides access to this data through their platform.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Social Services
Year: 2025 • Agency: Department of Social Services • Department: Public Engagement Unit
Year
2025
Agency
Department of Social Services
Department
Public Engagement Unit
Tool Name
Statara
Date First Use
2025/08
Updated
Tool was created in CY2025
Purpose Type
Data management
Computation Type
Matching
Autonomy
Informative
Frequency
Occasionally
Population Type
Individuals
Population Type Individual
New Yorkers who may be eligible for certain benefits
Population Type Other
NA
Website
https://statara.com/about-us/
Tool Desc
Statara leverages algorithmic matching and scoring to provide a commercial dataset of New York City residents and their demographic, contact, and other information. The Department of Social Services (DSS)/the Mayor's Public Engagement Unit (PEU) uses this data to create outreach lists when data on existing clients or from partner agencies is unavailable or insufficient to meet the scope of the outreach project.
Purpose Desc
In 2025, PEU used Statara data to drive outreach on several projects. For example, PEU used Statara to create lists of New Yorkers potentially eligible for Senior Citizen Rent Increase Exemption using factors like age and whether they were likely not a homeowner. For these projects, PEU uses Statara-provided addresses and/or phone information to conduct outreach.
Updated Desc
NA
Identifying Info
Training data; Input data; Output data
Data Training
Training data is part of vendor’s proprietary processes. The vendor receives no DSS/PEU data as part of training or input.
Data Input
Input data is part of vendor’s proprietary processes. The vendor receives no DSS/PEU data as part of training or input.
Data Output
The algorithmically derived data that PEU accesses is the output of proprietary algorithmic processes developed and operated by Statara. These algorithmic processes include matching multiple input datasets to determine residency, contact information, and demographics on New York City residents. Statara also includes a number of algorithmically determined likelihood scores, including scores for the likelihood that a household contains children under 18, etc. PEU receives a regularly updated data file from Statara for its use. PEU subsets and/or merges this file with other PEU data sources to drive outreach projects.
Vendor Name
Statara
Vendor Type
Procurement/Paid pilot, Professional services
Vendor Desc
Statara provides us with an updated consumer data file for New York City residents on a monthly basis.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Social Services
Year: 2025 • Agency: Department of Social Services • Department: Public Engagement Unit
Year
2025
Agency
Department of Social Services
Department
Public Engagement Unit
Tool Name
Tenant Harassment Risk Model (THRM)
Date First Use
2025/10
Updated
Tool was created in CY2025
Purpose Type
Resource allocation
Computation Type
Scoring
Autonomy
Informative
Frequency
Informs weekly outreach starting in October
Population Type
Property; Individuals
Population Type Individual
Tenants may or may not receive outreach due to the score of their building
Population Type Other
NA
Website
Not specified
Tool Desc
The Tenant Harassment Risk Model (THRM) was developed by Housing Preservation and Development (HPD) to identify buildings where tenants were more likely to experience harassment or other challenges with their landlord.
Purpose Desc
In collaboration with HPD, PEU uses the THRM model to identify buildings for proactive tenant outreach (door knocking) to identify individual issues as well as buildings that could benefit from tenant organizing through the HPD Partners in Preservation program.
Updated Desc
NA
Identifying Info
Training data; Input data; Output data
Data Training
The THRM uses data from multiple sources for training and input. Much of the data is from NYC Open Data, as well as agency partners and third parties such as Property Shark. The data includes variables such as violations, 311 requests/complaints, building financials, eviction filings. The methodology did not involve segmenting training data. All residential buildings were used in the analysis to develop the weights.
Data Input
The THRM uses data from multiple sources for training and input. Much of the data is from NYC Open Data, as well as agency partners and third parties such as Property Shark. The data includes variables such as violations, 311 requests/complaints, building financials, eviction filings. All residential buildings are included in input data.
Data Output
The THRM combines the output of multiple linear regressions to generate component scores focused on Construction, Eviction, Financial Distress, and Physical Distress (all on a 0-100 score) as well as an overall Tenant Harassment Risk Score (0-100). PEU received scores for residential buildings with at least 11 units and in which at least 50 percent of units are rent regulated.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Transportation
Year: 2025 • Agency: Department of Transportation • Department: Traffic Operations
Year
2025
Agency
Department of Transportation
Department
Traffic Operations
Tool Name
Midtown in Motion / Adaptive Control Decision Support System
Date First Use
2011/07
Updated
No
Purpose Type
Performance evaluation
Computation Type
Classification
Autonomy
Monitored
Frequency
Continuous
Population Type
Geographic space; Individuals
Population Type Individual
Travelers in the Manhattan midtown core
Population Type Other
NA
Website
Not specified
Tool Desc
Midtown in Motion (MIM) is a program that measures traffic congestion in the midtown core of Manhattan (from 1st to 9th Avenues and from 57th to 34th Streets, inclusive), using sensor data captured within and beyond the zone. This congestion classification (light, moderate, moderate-heavy, heavy) is used by the Adaptive Control Decision Support System (ACDSS) to choose the optimal signal timing along the avenues to reduce congestion and improve traffic flow for all modes of transportation.
Purpose Desc
The purpose of the tool is to improve congestion in the midtown core of Manhattan through responsive traffic signal management.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
N/A
Data Input
Motor vehicle travel times captured by Radio Frequency Identification (RFID) sensors.
Data Output
Recommended traffic signal plan.
Vendor Name
KLD Engineering, P.C.
Vendor Type
Procurement/Paid pilot
Vendor Desc
KLD Engineering, P.C. is the developer of ACDSS and currently handles the maintenance contract.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Youth and Community Development
Year: 2025 • Agency: Department of Youth and Community Development • Department: Communications
Year
2025
Agency
Department of Youth and Community Development
Department
Communications
Tool Name
Adobe Creative Suite
Date First Use
2025/05
Updated
Tool was created in CY2025
Purpose Type
Information presentation
Computation Type
Data transformation
Autonomy
Informative
Frequency
Weekly
Population Type
Individuals; Group, organization, or business
Population Type Individual
People in New York City
Population Type Other
NA
Website
Not specified
Tool Desc
Adobe Creative Cloud (CC) is a comprehensive collection of cloud software applications focused on graphic, video, and photography tools.
Purpose Desc
The creative suite has implemented AI tools in it's software. For Adobe Photoshop, you are able to use AI to spot-edit photos if you are trying to remove a logo on an article of clothing for example.
Updated Desc
NA
Identifying Info
Input data; Output data
Data Training
Training data is proprietary to the vendor.
Data Input
Photos, videos, graphics uploaded by a staff member to edit and transform into content for social media.
Data Output
Designed audio, visuals and graphics based on the designer who exports a finished asset.
Vendor Name
Adobe
Vendor Type
Off-the-shelf
Vendor Desc
Adobe developed and maintains the tool.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Youth and Community Development
Year: 2025 • Agency: Department of Youth and Community Development • Department: Communications
Year
2025
Agency
Department of Youth and Community Development
Department
Communications
Tool Name
Adobe Premier Pro
Date First Use
2025/04
Updated
Tool was created in CY2025
Purpose Type
Information presentation
Computation Type
Data transformation
Autonomy
Informative
Frequency
Weekly
Population Type
Individuals
Population Type Individual
People in New York City
Population Type Other
NA
Website
Not specified
Tool Desc
Adobe Premier Pro is a video editing software used to create post-event recap videos and highlight reels, including AI-generated captions.
Purpose Desc
Adobe Premier Pro uses AI to provide auto-subtitle captioning for videos.
Updated Desc
NA
Identifying Info
Input data; Output data
Data Training
Training data is proprietary to the vendor.
Data Input
Video files are uploaded to clip, cut, and edit into a final video.
Data Output
Exported video with music, subtitles, and professional editing.
Vendor Name
Adobe
Vendor Type
Off-the-shelf
Vendor Desc
Adobe developed and maintains the tool.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Youth and Community Development
Year: 2025 • Agency: Department of Youth and Community Development • Department: Communications
Year
2025
Agency
Department of Youth and Community Development
Department
Communications
Tool Name
ChatGPT
Date First Use
2024/04
Updated
No
Purpose Type
Information presentation
Computation Type
Data transformation
Autonomy
Informative
Frequency
Weekly
Population Type
Individuals; Group, organization, or business
Population Type Individual
People in New York City
Population Type Other
NA
Website
Not specified
Tool Desc
ChatGPT is an advanced AI language model that can understand and generate text based on the input it receives. It can assist with a wide range of tasks, including summarizing text, rephrase text to a lower reading level, and assist in providing content ideas for social media.
Purpose Desc
DYCD's Communication Unit have used ChatGPT to complete the following tasks: assist in summarizing text to create more concise captions, helping brainstorm DYCD related content ideas for social media, and taking current text and rewriting it to any necessary reading level.
Updated Desc
NA
Identifying Info
Input data
Data Training
Training data is proprietary to OpenAI.
Data Input
DYCD staff enter text prompts related to creating and editing social media content.
Data Output
Text response to the user prompt, which is reviewed and adapted by staff prior to publication.
Vendor Name
OpenAI
Vendor Type
Off-the-shelf
Vendor Desc
ChatGPT is a freely available tool created and maintained by OpenAI.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Youth and Community Development
Year: 2025 • Agency: Department of Youth and Community Development • Department: Information Technology - Application Development
Year
2025
Agency
Department of Youth and Community Development
Department
Information Technology - Application Development
Tool Name
DYCD Virtual Assistant Chatbot
Date First Use
2020/09
Updated
Yes
Purpose Type
Information presentation
Computation Type
Matching
Autonomy
Fully autonomous
Frequency
Averaging about requests 2100 a day
Population Type
Individuals
Population Type Individual
People in New York City
Population Type Other
NA
Website
https://www.nyc.gov/dycd
Tool Desc
The DYCD Virtual Assistant Chatbot provides support to the Community Connect team by providing predefined answers to questions asked to the Chatbot by the public. This tool enables the public to quickly get answers to the most commonly asked questions on DYCD programs while reducing the number of requests to the Community Connect team on basic questions. The tool allows for the Community Connect team spend more time supporting the public on more nuanced scenarios and complex questions.
Purpose Desc
Information received is used to understand public questions and public interest in programs offered to produce the matching predefined response relevant to the question.
Updated Desc
The DYCD Virtual Assistant Chatbot was upgraded in 2025 to use the Microsoft Azure Language, a cloud-based service that provides natural language processing features for understanding and analyzing text to help build intelligent applications. In particular, the chatbot uses the Custom Question Answering and Conventional Language Understanding components.
Identifying Info
Input data
Data Training
Staff worked to capture and maintain information regarding the programs in a knowledge base.
Data Input
Questions are input to chat bot from user in general text format.
Data Output
The Conversational Language Understanding (CLU) model is then used to capture the users intent, which in turn is directed to our Custom Question Answering (CQA) model, that searches the knowledge base, ranks the possible answers and returns the best match.
Vendor Name
Microsoft
Vendor Type
Procurement/Paid pilot
Vendor Desc
Microsoft Azure provides the software for implementing the chatbot.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Department of Youth and Community Development
Year: 2025 • Agency: Department of Youth and Community Development • Department: Communications
Year
2025
Agency
Department of Youth and Community Development
Department
Communications
Tool Name
OpusClip
Date First Use
2024/06
Updated
Yes
Purpose Type
Information presentation
Computation Type
Ranking
Autonomy
Informative
Frequency
Biweekly
Population Type
Individuals; Group, organization, or business
Population Type Individual
Any viewer watching the content. Our targeted audience is New Yorkers
Population Type Other
NA
Website
Not specified
Tool Desc
OpusClip uses their data and AI to analyze video in the context of current social and marketing trends. It detects the most compelling, shareable, and emotionally impactful moments within the video that is uploaded.
Purpose Desc
Primarily used for our DYCD Next Gen Podcast. DYCD will record the podcast and use OpusClip to find moments in the video that may bring the most engagement when we post on our social media platforms.
Updated Desc
Tool is a web platform and updates regularly.
Identifying Info
Input data; Output data
Data Training
Training data is proprietary to the vendor.
Data Input
Finalized video meant for public viewing.
Data Output
The platform analyzes an uploaded video and provides the time stamps and clips of the original video that have a good chance of creating engagement with the viewer to comment, share, and like. We can then export the selected clips.
Vendor Name
OpusClip
Vendor Type
Off-the-shelf
Vendor Desc
OpusClip develops and maintains the product.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Fire Department
Year: 2025 • Agency: Fire Department • Department: Management, Analysis, and Planning
Year
2025
Agency
Fire Department
Department
Management, Analysis, and Planning
Tool Name
Emergency Medical Services (EMS) Hospital Suggestion Algorithm
Date First Use
2020/12
Updated
No
Purpose Type
Resource allocation
Computation Type
Ranking
Autonomy
Informative
Frequency
Daily
Population Type
Individuals; Geographic space
Population Type Individual
Patients
Population Type Other
NA
Website
https://doi.org/10.1016/j.scs.2022.104104
Tool Desc
The Emergency Medical Services (EMS) Hospital Suggestion Algorithm is used to determine the closest, appropriate hospital to the incident location based on the medical needs of a patient requiring transport.
Purpose Desc
The algorithm computes a list of hospitals in order of closest to furthest in time for each medical condition category as currently established. (For example, there is a list of hospitals computed in order of closest in time for all hospitals that accept General Emergency Department patients and for all hospitals that accept special conditions, such as burns.) Depending on the medical needs category of the patient, the algorithm produces a pre-determined list of hospitals which is based on the location of the patient and then made available to the crew as a list of “closest, most appropriate hospitals.”
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
The EMS Hospital Suggestion algorithm relies on automatic vehicle location data from ambulances transporting patients to hospitals between 2018 and 2019 to calibrate a network analysis model that derives incident to hospital transport times. The order of suggested hospitals are then compared with five years of historical EMS hospital transport data from before the COVID-19 pandemic (2015-2019) to validate and correct the network model.
Data Input
The inputs for the algorithm include the location and medical call type of the patient.
Data Output
The algorithm outputs the closest, most appropriate hospitals.
Vendor Name
None
Vendor Type
NA
Vendor Desc
NA
Data 2022
NA
Vendor
NA
Analysis Type
NA
Fire Department
Year: 2025 • Agency: Fire Department • Department: Management Analysis and Planning
Year
2025
Agency
Fire Department
Department
Management Analysis and Planning
Tool Name
Emergency Medical Services (EMS) Unit Suggestion Algorithm
Date First Use
2007/03
Updated
Yes
Purpose Type
Resource allocation
Computation Type
Ranking
Autonomy
Informative
Frequency
Daily
Population Type
Geographic space; Individuals
Population Type Individual
individuals calling 911 for emergency medical services
Population Type Other
NA
Website
Not specified
Tool Desc
The Emergency Medical Services (EMS) Unit Suggestion Algorithm is used to determine which order of geographic regions (known as atoms) to search in order for the EMSCAD system to select an appropriate EMS unit for dispatch to an incident.
Purpose Desc
The algorithm computes a list of geographic regions (known as atoms) in order of closest to furthest in travel time for each atom in the city. This list of ordered atoms is the output of an algorithm that relies on a calibrated network model to derive travel time estimates. The output is an excel file which is converted into an EMSCAD-compatible file and loaded into the system for real-time unit selection capabilities. Starting in September 2025, The file was generated and implemented as a 24/7 source file, meaning the recommended search order is not currently varying by time of day. In September 2025, the model was updated to accept time of day as an input.
Updated Desc
Starting in September 2025, input data includes time of day to make computations. The output of this process now includes outputs for different times of day and day of week.
Identifying Info
No identifying information is collected or disclosed
Data Training
The EMS Unit Suggestion algorithm relies on historical EMSCAD trip time data which is used to calibrate a network analysis model which derives atom-to-atom transport times.
Data Input
The input for the algorithm is a geographic location, time of day (as of September 2025), and historical emergency unit response time data.
Data Output
The algorithm outputs a recommended EMS unit for dispatch.
Vendor Name
Deccan International
Vendor Type
Professional services
Vendor Desc
This algorithm and the resulting output file that is used in our EMSCAD system to suggest atom order for unit search is currently provided by a vendor, Deccan International.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Fire Department
Year: 2025 • Agency: Fire Department • Department: Bureau of Fire Investigations
Year
2025
Agency
Fire Department
Department
Bureau of Fire Investigations
Tool Name
Facial Recognition Technology
Date First Use
2022/12
Updated
No
Purpose Type
Information presentation
Computation Type
Classification
Autonomy
Informative
Frequency
Not specified
Population Type
Individuals
Population Type Individual
NA
Population Type Other
NA
Website
Not specified
Tool Desc
The tool analyzes an uploaded image or video and compares the image to lawfully possessed images to generate a pool of possible matches. Any matches serve only as investigative leads and Bureau of Fire Investigations personnel conduct an additional investigation. The match alone is not treated as a positive identification.
Purpose Desc
The purpose of the tool is to assist Fire Department investigations into arson and other criminal activities that fall within the agency's jurisdiction.
Updated Desc
NA
Identifying Info
Training data; Input data; Output data
Data Training
Publicly available open-source social media data.
Data Input
Images of individuals of interest and images from social media or the internet.
Data Output
The tool generates possible image matches of an object or individual.
Vendor Name
Not disclosable
Vendor Type
Off-the-shelf
Vendor Desc
Vendor provides ongoing technical support. Confidentiality agreements are in place with the vendor.
Data 2022
NA
Vendor
NA
Analysis Type
NA
Fire Department
Year: 2025 • Agency: Fire Department • Department: Bureau of Fire Investigations
Year
2025
Agency
Fire Department
Department
Bureau of Fire Investigations
Tool Name
Fire Brush Fire Detection Camera
Date First Use
2025/09
Updated
Tool was created in CY2025
Purpose Type
Risk management
Computation Type
Classification
Autonomy
Informative
Frequency
Not specified
Population Type
Property; Geographic space
Population Type Individual
NA
Population Type Other
NA
Website
Not specified
Tool Desc
This tool is an early warning detection system that uses solar-powered cameras that send data to a secure AI server. The server analyzes the images using artificial intelligence to detect signs of smoke or fire.
Purpose Desc
If smoke or fire is detected on the camera, an alert is sent to the Bureau of Fire Investigations and the Fire Department Operation Center to view the camera and determine if an emergency response is required.
Updated Desc
NA
Identifying Info
No identifying information is collected or disclosed
Data Training
Reference images of smoke and fire-related objects.
Data Input
Images of treetops from FDNY cameras in City parks or other City-owned locations.
Data Output
Email alert and images.
Vendor Name
Claro AI
Vendor Type
Off-the-shelf, Procurement/Paid pilot
Vendor Desc
Vendor provides ongoing technical support.
Data 2022
NA
Vendor
NA
Analysis Type
NA
NYC Data AI • Public-data exploration interface