AUT Centre for Social Data Analytics (CSDA)

What AUT Centre for Social Data Analytics (CSDA) does

CSDA (Centre for Social Data Analytics) is a university research centre founded in 2016 within Auckland University of Technology's School of Social Sciences and Public Policy, with co-located operations in Brisbane, Australia. It builds and deploys machine learning-based predictive risk modelling (PRM) tools that operate on agencies' existing linked administrative data to support high-stakes triage decisions in social services — primarily child welfare hotline screening, child welfare investigation prioritisation, and housing/homelessness assistance prioritisation — with additional work in labour market nowcasting and workforce wellbeing. Its core products include the Allegheny Family Screening Tool (AFST, 2016 — the world's first child welfare PRM, independently validated by Stanford), the Allegheny Housing Assessment (AHA, 2020), the Housing Assistance Screening Tools (HAST and HAST(DW)), the Douglas County Decision Aide, the Hello Baby Program PRM, the LA County Risk Stratification Model, and the UR-NOW unemployment nowcasting tool. Research artefacts such as the Sibyl explainability interface and the Proficiente lifelong-learning framework underpin the production systems.

CSDA generates revenue through bespoke multi-year research contracts and service agreements with public sector agencies — chiefly US county and state Departments of Human Services (Allegheny, Douglas, Larimer, Northampton, Los Angeles Counties) and New Zealand ministries (Ministry of Social Development, Oranga Tamariki) — supplemented by competitive research grants (NZ Royal Society Marsden Fund, Children and Families Research Fund). Go-to-market is sales-led, with a dedicated Business Development Manager engaging agencies directly, reinforced by thought leadership through peer-reviewed publications (JAMA Pediatrics, AJPH, Pediatrics), international conference presentations (Nobel Conference 57, Bloomberg D4GX, FAT*), and significant earned media coverage (New York Times Magazine, Netflix's 'The Trials of Gabriel Fernandez', Nature). All deployments are subject to independent ethical and data science reviews. The centre employs ~11 staff across New Zealand, Australia, and the United States and is led by Director and co-founder Rhema Vaithianathan, with Deputy Director Dr Nina Anchugina.

Name

AUT Centre for Social Data Analytics (CSDA)

Legal name

Centre for Social Data Analytics (CSDA)

Website

https://csda.aut.ac.nz

Company type

Private

Founded year

2016

Operating status

Operating

Headcount range

11–50 employees

Short description

CSDA builds and deploys machine learning predictive risk modelling tools for government social services — child welfare hotline screening, housing prioritisation, and unemployment nowcasting — primarily for US counties and New Zealand ministries, using existing administrative data and independent ethical review.

Ownership category

akta.pro rank

Product category

Predictive Analytics for Social Services

NAICS

Individual and Family Services (6241), Administration of Housing Programs, Urban Planning, and Community Development (925), Research and Development in the Social Sciences and Humanities (541720)

SIC

Services-Social Services (8300), Services-Computer Programming, Data Processing, Etc. (7370), Services-Management Consulting Services (8742)

akta.pro primary industry

State/Provincial Human Services & Benefits Administration (Welfare, Disability, Child & Family) (BPAIABAI)

Keywords

• Predictive risk modelling
• Social data analytics
• Child welfare screening
• Homelessness triage tools
• Machine learning for government

Headquarters

HQ city

Auckland

HQ country

New Zealand

HQ region

Oceania

Offices2 records

Markets served

GTM type

B2B

Offering type

Software

Cost components

Personnel, Technology or R&D, Operations, Marketing or Sales, Others

Revenue model

1. Government research contracts and service agreements: CSDA operates as a university research centre generating revenue through commissioned research contracts with government agencies, primarily in the US (county and state governments), New Zealand, and Australia. Contracts typically cover the development, validation, deployment, and training for predictive analytics tools. Examples include multi-year contracts with Northampton County, PA and ongoing collaborations with Allegheny County Department of Human Services.

Go-to-market motion2 records

Distribution channels1 record

Marketing channels7 records

Core offering

CSDA is a university research centre at Auckland University of Technology that designs, builds, validates and deploys predictive analytics tools for government social-service agencies. Its core deliverables are machine-learning risk models (e.g., the Allegheny Family Screening Tool, Allegheny Housing Assessment, Housing Assistance Screening Tools, Douglas County Decision Aide, Hello Baby Program PRM, LA County Risk Stratification Model, and UR-NOW unemployment nowcasting) built on existing administrative data to support high-stakes decisions in child welfare, homelessness and labour-market monitoring. Engagements are delivered as bespoke, multi-year research and service contracts with public-sector clients in the US, Australia and New Zealand.

Product overview

CSDA offers a portfolio of predictive analytics tools for social services, primarily focused on child welfare and homelessness prioritisation in US jurisdictions. The core product line includes: the Allegheny Family Screening Tool (AFST) for child maltreatment hotline screening decisions; the Allegheny Housing Assessment (AHA) tool and Housing Assistance Screening Tools (HAST/HAST(DW)) for housing assistance prioritisation; the Douglas County Decision Aide (DCDA) for child welfare referrals; the Hello Baby Program PRM for proactive child harm prevention; and the Risk Stratification Model for LA County investigations. Supporting tools include UR-NOW for unemployment nowcasting, Sibyl for ML explainability in child welfare contexts, and the MyDay survey tool for workforce wellbeing research. All tools share a common approach of using machine learning on administrative data to generate risk scores that support human decision-making while emphasising transparency, fairness, and ethical oversight.

Differentiator

Problem solved

Functional benefit

Products and services

• Allegheny Family Screening Tool (AFST) World-first child welfare predictive risk modelling tool that uses rich administrative data to generate screening scores for incoming calls alleging child maltreatment and neglect, helping call screeners decide whether to open an investigation for child protection. Deployed in Allegheny County, PA since 2016.
• Allegheny Housing Assessment (AHA) Tool Predictive risk modelling tool deployed by Allegheny County in September 2020 that helps prioritise clients for longer-term housing programs. Uses machine learning trained on local administrative data to identify clients at highest risk of housing-related adversities including mental illness, jail time and health emergencies.
• Housing Assistance Screening Tool (HAST) Triaging tool using client survey and HMIS (Homeless Management Information System) data to support prioritisation of highest-need housing clients in US jurisdictions. Accurately identifies at-risk clients while respecting privacy and minimising intrusive questions.
• Housing Assistance Screening Tool Data Warehouse (HAST(DW)) Triaging tool using administrative data to support prioritisation of highest-need housing clients. Eliminates reliance on self-reporting and can flag risk in Black and African American clients more accurately than theoretically validated methods.
• Douglas County Decision Aide (DCDA) Child welfare decision aid tool that produces risk scores (1-20) for incoming child maltreatment referrals, helping County staff decide whether referrals should be investigated. Deployed in early 2019 in Douglas County, CO, with an independent randomised controlled trial evaluation by Cornell University.
• Hello Baby Program Predictive Risk Model Screening tool developed for Allegheny County's Hello Baby Program to identify families with highest needs for proactive child harm prevention. Uses existing administrative data to stratify families by risk of adverse outcomes across differentiated service levels (Universal, Family Support, Priority), targeting children aged zero to three.
• UR-NOW (Unemployment Rate Nowcast) Predictive model developed in collaboration with the NZ Ministry of Social Development that uses administrative data about benefit receipt and filled jobs to predict monthly unemployment rates, supplementing official quarterly HLFS figures with more timely information. The Q3 2025 consensus result exactly predicted the official rate of 5.3%.
• LA County Risk Stratification Model Predictive risk model for child welfare investigations piloted in three Los Angeles County DCFS regional offices (Belvedere, Lancaster, Santa Fe Springs). Trained to assess children's risk of long-arc adverse outcomes and generates daily scores for complex-risk protocols, investigation overview reports, and racial equity feedback loops.
• Northampton County Decision Aid Tool Custom child welfare decision aid tool built under a three-year contract for the Children, Youth and Families Division of the Northampton County Department of Human Services. Integrates and analyses hundreds of child welfare data elements to predict the likelihood of future child welfare involvement.
• MyDay Survey Tool Purpose-built web-based survey tool that collects anonymous data about the wellbeing of workforces. Used in studies including Health Education England Wessex to understand task-related factors driving trainee doctor wellbeing, involving 565 trainee doctors.

Quantifiable outcome

• Children with high AFST scores are 9x more likely to experience abusive injury hospital encounter and 10x more likely to experience suicide or self-inflicted harm compared to children with low risk scores.
• +4 more outcomes

Named customers8 records

Segments4 records

Ideal customer profiles4 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability13 records

Feature7 records

Partnerships

25 partnerships are on record, tiered core, major and minor.

• Allegheny County Department of Human Services (USA)coreStrategic or Co-development Partner · 1 January 2014Long-standing strategic partnership for the development and deployment of child welfare and housing predictive analytics tools. Includes the Allegheny Family Screening Tool (deployed 2016), Allegheny Housing Assessment (deployed 2020), and Hello Baby Program. The county has commissioned independent ethical reviews and evaluations.
• Children's Data Network, University of Southern California (USA)coreStrategic or Co-development PartnerCollaborates with CSDA on child welfare predictive risk modelling projects. Emily Putnam-Hornstein serves as Director of the Children's Data Network and is a co-author on multiple CSDA publications and projects including the AFST and Hello Baby Program.
• Carnegie Mellon University (USA)coreStrategic or Co-development PartnerAcademic partner contributing to the development of the Allegheny Family Screening Tool methodology and algorithm analysis. Alexandra Chouldechova from CMU co-authored the FAT* best paper award paper.
• Oranga Tamariki (New Zealand)coreStrategic or Co-development PartnerNew Zealand's Ministry for Children (Oranga Tamariki) collaborates on research into childhood adversities, school readiness, and child welfare data analytics. Co-funded research through the Children and Families Research Fund.
• Ministry of Social Development (New Zealand)coreStrategic or Co-development PartnerCollaboration on the UR-NOW unemployment nowcasting tool, led by AUT Professor Tim Maloney who also serves as Chief Economist at the Ministry. The Ministry's Principal Advisor David Rea co-authored the working paper.
• Stats NZmajorStrategic or Co-development PartnerCo-presented the 2018 'Algorithms in the Public Sector' conference with CSDA and The University of Auckland. Stats NZ is New Zealand's national statistical office.
• Institute for Social Science Research (ISSR), University of Queensland (Australia)coreStrategic or Co-development PartnerFive-year collaboration to establish a social data analytics capability at UQ. CSDA Director Rhema Vaithianathan holds a partial appointment as Professor of Social Data Analytics at ISSR. The vision is to establish a fully staffed CSDA site at UQ by 2025.
• Health Education England, Working Across Wessex (UK)majorStrategic or Co-development PartnerPartner on the MyDay Survey project studying trainee doctor wellbeing at work, involving 565 trainee doctors. Also involves University of Southampton, City University of London, and University of Zurich.
• The University of Auckland (New Zealand)majorStrategic or Co-development PartnerCo-presented the 2018 'Algorithms in the Public Sector' conference and collaborates on ethics of social policy predictive risk modelling research (NZ Royal Society Marsden Fund Grant, PI Tim Dare, A.I. Rhema Vaithianathan).
• Children's Data Network (USA)coreStrategic or Co-development PartnerLed by Emily Putnam-Hornstein and Jacquelyn McCroskey, the CDN led a risk stratification project for Los Angeles County Department of Children and Family Services, with CSDA building the predictive model.
• Actionable Intelligence for Social Policy (AISP), University of Pennsylvania (USA)majorStrategic or Co-development PartnerCollaboration on housing assistance tools. Professor Dennis Culhane chairs the AISP initiative. Jointly hosted a September 2021 webinar on Housing Assistance Screening Tools with the National Alliance to End Homelessness.
• National Alliance to End Homelessness (USA)majorStrategic or Co-development PartnerJointly hosted a September 2021 webinar on Housing Assistance Screening Tools with CSDA and AISP. The Alliance is the largest convening organisation for homelessness service providers and advocates in the United States.
• Bloomberg Philanthropies / Bloomberg Data for Good Exchange (USA)majorStrategic or Co-development PartnerCSDA researchers presented the Allegheny Homelessness Tool at the Bloomberg Data for Good Exchange (D4GX) conference in New York in September 2019. D4GX is an annual event bringing together data scientists with non-profits and government agencies for social good.
• Ministry of Social Development (New Zealand)majorStrategic or Co-development PartnerOne of the funders/supporters of the 2018 'Algorithms in the Public Sector' conference. Also collaborates on child welfare and childhood adversities research through funding rounds.
• ACC (New Zealand)minorStrategic or Co-development PartnerProvided funding and support for the 2018 'Algorithms in the Public Sector' conference.
• National Alliance to End Homelessness (USA)majorGTM or Marketing PartnerHosted CSDA's September 2021 webinar on housing triage tools at the National Virtual Conference on Ending Homelessness, reaching nearly 200 representatives of US health and human services agencies.
• Institute for Social Science Research (ISSR), University of Queensland (Australia)coreStrategic or Co-development PartnerCo-authored the June 2020 working paper 'Digital Contact Tracing for COVID-19: A Primer for Policymakers' with CSDA.
• Growing Up in New Zealand (GUINZ) longitudinal studycoreStrategic or Co-development PartnerLed by the University of Auckland, this longitudinal study of 6,800+ children provides data for CSDA research on childhood adversities and school readiness. CSDA received government funding through the Children and Families Research Fund to use GUINZ data.
• Tim Maloney (Co-founder)coreOthersProfessor Tim Maloney co-founded CSDA with Rhema Vaithianathan in 2016 and served as co-director until July 2019. He continues as Head of the School of Social Sciences and Public Policy at AUT and contributed to the UR-NOW tool as Chief Economist at MSD.
• Adolfo Ibáñez University (UAI), ChileminorStrategic or Co-development PartnerProfessor Rhema Vaithianathan gave a keynote address on child welfare predictive risk models at UAI in Santiago, Chile in June 2018.
• Northampton County Department of Human Services (USA)coreStrategic or Co-development PartnerContracted in March 2021 to deliver a child welfare decision aid tool over three years. The tool will integrate and analyse hundreds of child welfare data elements to predict likelihood of future child welfare involvement.
• Cornell University (USA)majorStrategic or Co-development PartnerProfessor Chris Wildeman and Associate Professor Maria Fitzpatrick at Cornell are conducting an independent randomised controlled trial evaluation of the Douglas County Decision Aide.
• Warren and Mahoney Architects (New Zealand)minorStrategic or Co-development PartnerPilot MyDay study extending the wellbeing survey tool to measure the effect of physical workspace environments on staff wellbeing. Pilot project to test a new dimension of the MyDay tool.
• Queensland Health / Paediatric Intensive Care Units (Australia)minorStrategic or Co-development PartnerDeveloping a machine learning tool to identify children admitted to paediatric intensive care units in Queensland who are at risk of poor future educational outcomes. Also developing tools for child and youth mental health services and an Aboriginal and Torres Strait Islander primary care clinic.
• University of Los Andes (Colombia)minorStrategic or Co-development PartnerDeveloped and delivered a three-week intensive course 'Machine Learning for Decision-Making Support in Public Sector High-Stakes Problems' for master's students via Zoom in July 2021.

Scale indicators5 records

Recent moves8 records

Expansion highlights6 records

Broad incumbents

• MDRC: Major US social policy research organization with programs in family wellbeing, housing, and poverty. Overlaps with CSDA's evidence-building and program evaluation mandate.
• Palantir Technologies: Commercial data analytics provider serving public sector agencies with predictive analytics platforms. Represents the strategic competitive threat if CSDA attempts to scale commercially.
• Urban Institute: US policy think tank with Health Policy Center, housing, and data science capabilities. Competes for government-funded research mandates in social services.
• RAND Corporation: Global policy research institution with child welfare and algorithmic fairness research. Competes for high-profile public-sector studies and produces comparable thought leadership.
• Mathematica Policy Research: Large US policy research and data analytics firm with deep child welfare, housing, and human services evaluation work. Competes for the same government contract dollars as CSDA but at much greater scale.

Direct peers

• Chapin Hall at the University of Chicago: University-affiliated research center producing data-driven evidence on child welfare, child welfare informatics, and poverty. Operates adjacent to CSDA in the academic child-welfare research ecosystem.
• Children's Data Network (USC): University of Southern California research center co-developing child welfare predictive risk models with CSDA, including the LA County Risk Stratification Model. Closest direct peer in child welfare PRM research.
• Priority Metrics Group (now part of CSCI) / OrgCode Consulting: Vendor of the VI-SPDAT (the housing assessment tool CSDA's AHA replaced). Most direct competitor in homelessness triage tooling for US human services agencies.
• Actionable Intelligence for Social Policy (AISP), University of Pennsylvania: UPenn-based research initiative on integrated administrative data for social policy. Co-developed CSDA's Housing Assistance Screening Tools and is a core academic partner.

Others

• Oranga Tamariki (NZ Ministry for Children): New Zealand government agency that is both a research partner with CSDA and a potential client/buyer of CSDA-developed tools. Listed as an indirect ecosystem comparable.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks6 records

Key highlights7 records

Customer concentration

• Xx.com/autcsda
• LinkedInlinkedin.com/company/centre-for-social-data-analytics

Revenue estimate

Valuation estimate

Number of profiles

Profiles6 records

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about AUT Centre for Social Data Analytics (CSDA)