Innovation

The Center for Innovation through Data Intelligence (CIDI) is a research and policy center that conducts citywide, interagency research to improve services in New York City. Its vision is to “make data come alive to inspire change.” It fosters collaboration mostly among Health and Human Service agencies, which serve a wide range of vulnerable populations and are particularly attentive to privacy and data security. CIDI researchers analyze and visualize data to identify service gaps, answer policy questions, and examine the effectiveness of city programs and services. Its notable projects include the Disparity Report, which integrates racial-equity data from multiple life stages and experiences into one tool so that government agencies and community partners can understand and address the disparities faced by young people of color.

Democratic Challenge

Big social challenges like poverty, inequity, misogyny and discrimination do not fall within the purview of a single government agency. Their causes and effects cut across all sectors and services. But in bureaucracies, each department typically has its own siloed information, making it hard for policymakers to see and understand how complex phenomena are affecting city services and residents.

How Did They Do It?

Before CIDI came into existence, New York City had been spending a lot of time overcoming policy and legal barriers in order to extract, link and conduct statistical analysis on data that is collected across government departments. These long-running efforts laid the foundation for CIDI, which was created in 2011 and functions as an integrated data system for health and social services. CIDI created a legal and policy process for Health and Human Service agencies to identify the right questions that could be answered by administrative data, and it set out to build upon cross-agency collaboration efforts by using interagency data to guide policy and practice.

CIDI is located in the mayor’s office and reports to the Deputy Mayor for Health and Human Services (HHS,) who oversees 11 city agencies. CIDI works most closely with these agencies, but maintains relationships with other departments (such as Housing and Education) as well. It also participates in multi-city research initiatives such as Actionable Intelligence for Social Policy.

Within the umbrella of HHS, several agencies have their own in-house data analysis teams. These agency teams tend to focus on single-issue studies using agency-specific data. CIDI works with these teams to conduct inter-agency, multi-issue studies that aggregate critical data from different sources to inform decision making. “We build upon the agencies’ knowledge and bring together the cross-system collaboration,” Director Maryanne Schretzman said. CIDI’s location inside government also gives it the ability to take on multi-year research projects which use cross-agency outcome data. In one such example, they evaluated a supportive housing program and included outcomes such as jail and homeless shelter stays.

CIDI follows a consistent process to ensure that its studies are tailored to the needs of the individuals, families and children involved in human service agencies. The research agenda of CIDI is driven by the Deputy Mayor for Health and Human Services along with the Commissioners and leaders of City agencies. The typical process is:

1. The Deputy Mayor will present a question.
2. CIDI convenes a working group, typically with social workers and data scientists from the relevant agencies.
3. The working group develops an operational question and operational plan. In this way context and trust are assured throughout the process.
4. The Deputy Mayor approves the operational plan.
5. CIDI’s protocol includes contacting the lawyers at relevant agencies to assure that confidentiality requirements for requested data are met. Agencies have an incentive to approve CIDI requests because the Deputy Mayor approved the research plan, and because participating may result in new resources.
6. Participating agencies use CIDI’s Data Transfer Protocol to share information quickly.
7. The working group participates in the analysis to ensure the appropriate interpretation of the findings and the feasibility of action steps.

With funding from the Arnold Foundation, CIDI has sent teams of staff to learn machine learning, python, and other big-data techniques. Schretzman said this brought new skills in-house but was also a good professional-development opportunity for people who may be experts in their own agency data system but are buried in reporting and regulatory requirements, leaving little time for new ways to think about and analyze data.

In sum, CIDI is able to add value to the work of agencies because it has:

1. Developed Trust. Line departments and their staff have a lot on their plate and through collaboration with CIDI, interagency work can be done more efficiently.
2. Authority. CIDI derives its mandate directly from the Deputy Mayor of Health and Human Services.
3. License. There are very few restrictions on where CIDI can go to get the best information.
4. External Expertise. CIDI can tap and use a variety of networks and consultants to find the best method for uncovering solutions.

How Is It Going?

• CIDI’s landmark Disparity Report was well-received by the City Council and helped persuade policymakers of the value of integrated data analysis. The Report provides insights into racial disparities in four main areas: education; economic security and mobility; health; and community and personal safety. The project was originally commissioned by the Young Men’s Initiative (YMI), a public-private partnership that seeks to address increasing disparities among black and Latino men ages 16-24. CIDI then expanded the methodology to understand racial disparities for both men and women over time, integrating data from multiple life stages and experiences into one tool.
• CIDI’s centralized system for multi-agency data analysis can be helpful in crisis situations. During Superstorm Sandy, CIDI used data collected by the National Guard to disseminate reports and help different agencies meet residents’ needs. CIDI also generated maps of residential units lacking heat, electricity, or phone communication, to assist hurricane-relief efforts.
• CIDI conducted a study which developed a typology of six groups of youth based on their service use and housing outcomes for the three years after they exit homeless services or foster care. It combined data from a number of systems to identify comprehensive service use profiles for discrete groups of youth. Utilizing this approach allowed the analysis to be multi-dimensional and represent multiple service needs of youth in each group, while also recognizing overarching differences among groups. In order to be actionable and to further demonstrate the distinction among groups, the study helped inform which groups of youth may need more services or interventions.

Considerations

• Ethical Data Sharing and Use. The Actionable Intelligence for Social Policy (AISP) initiative at the University of Pennsylvania has created a governance manual for setting up an ethical, effective data sharing system. Concerns around data integrity and personal privacy are often top of mind, especially as data integration creates more and more robust profiles of each person’s identity and daily routines. There are also emerging concerns about how predictive tools can inhibit equity and accountability. But AISP argues that such concerns should not stop data sharing completely, because governments have an ethical obligation to respectfully share and use the information they have in order to spur innovation and improve lives.