Homeless Predictive Data

Arizona Housing Analytics Collaborative (AzHAC)

“Direct Analytics to Decision-making and Care Coordination”

AzHAC is a tri-university collaborative among Arizona’s three public universities: Arizona State University, the University of Arizona, and Northern Arizona University that uses data collection, analysis, and dissemination to support policy and practice communities in preventing and intervening in homelessness across the state. With funding from the Garcia Family Foundation, Amazon Web Services’ Health Equity Initiative, and the U.S. Department of Housing and Urban Development (HUD), this initiative uses data analytics, predictive modeling, and qualitative inquiry to inform decision-making and care coordination. 

AzHAC’s work has multiple aims:

  1. Identify the health and psychosocial characteristics most likely to produce homelessness and result in successful exits from homelessness, including identifying characteristics and costs associated with high utilizers of HMIS services;
    Sample Analyses: High utilizer analysis, cost analysis, returns to homelessness; economic security program use among HMIS service users, linked 211-HMIS analysis
  2. Identify which programs, interventions, and assessment measures are most effective in producing positive outcomes for individuals experiencing homelessness;
    Sample Analyses: Prevention evaluation, housing program effectiveness, sequencing analysis, system dynamics analysis, comparative assessment of vulnerability measures
  3. Build and test predictive models to identify risk factors to prevent homelessness as well as interventions to reduce homelessness;
    Sample Analyses: Predictive models on who is most likely to successfully exit homeless; and on upstream indicators most likely to lead to homelessness and to eviction
  4. Build a social determinants of health (SDOH) housing assessment tool to identify characteristics and services among communities that effectively keep their residents housed;
    Sample Analyses: Identifying blocks around the state that have greatest/fewest characteristics related to housing stability/instability, evictions, and use of outreach services
  5. Prioritize analytical work around homelessness for DWEL-AZ, AHCCCS, and ADOH. 
    Sample Analyses: For MAG: how does MAP compare with VI-SPDAT? Who are the high utilizers of HMIS services? Who is most likely to return to homelessness?
Homeless Management Information System (HMIS)

Homeless Management Information System (HMIS)

Using combinations of data sources including Homeless Management Information System (HMIS) data, Arizona 2-1-1 data, state benefits data (e.g., food stamps, cash assistance, unemployment), city resources’ data (e.g., emergency rental assistance, utility payments and calls for assistance), court data (e.g., eviction records), education data, and criminal justice data, AzHAC’s work can better identify markers of housing insecurity and allow us to be proactive in engaging consumers with optimal intervention(s).