Boston CoC System Exploration App
Open app (Credentials required)
Purpose / Audience
This repo houses a web application that displays core metrics about the Boston Homelessness Continuum of Care that can be filtered across a variety of demographics and subpopulations.
The goal of this web app is to allow and empower the Supportive Housing Division to explore and ask questions of its HMIS data. The app provides answers to broad system performance questions like “is the average length of stay decreasing?” but also allows users to dig deeper and uncover insights from the data, like “White veterans are older and have shorter lengths of stay than veterans of color” (no clue if that’s true - just an example). The baseline measures displayed in the app are an adaption/expansion of HUD’s System Performance Measures. The Supportive Housing Division collaborated to workshop and tailor these measures to be more relevant to the goals of our CoC.
Data
The data used to power this app was pulled from the City of Boston’s HMIS. R scripts cleaned, reshaped, and deidentified HMIS data into key reporting tables, which were then written to a Postgres backend.
The app operates off the following key reporting tables:
Census
Number of individuals in Emergency Shelter / Street Outreach on a given night, by Gender, Race, Ethnicity, Household Status, and Veteran Status
Inflow
Number of individuals experiencing their first-ever stay in an Emergency Shelter / Street Outreach in a given year, by Gender, Race, Ethnicity, Household Status, and Veteran Status
Exits to Permanent Housing
Number of individuals exiting to a permanent housing destination in a given year, by Gender, Race, Ethnicity, Household Status, and Veteran Status
Returns to Shelter
Number of individuals returning to Emergency Shelter or Street Outreach from a prior exit to a permanent housing destination in a given year, by Gender, Race, Ethnicity, Household Status, and Veteran Status
The tables above all have the following structure, with ‘count’ corresponding to the appropriate metric in each table:
| year | race | ethnicity | gender | veteranstatus | householdtype | count |
|---|---|---|---|---|---|---|
Length of Stay
Average length of stay in Emergency Shelter / Street Outreach in a given year, by Gender, Race, Ethnicity, Household Status, and Veteran Status. This table has a slightly different structure so that we can take a weighted average length of stay depending which demographic filters are selected:
| year | race | ethnicity | gender | veteranstatus | householdtype | numclients | avglos |
|---|---|---|---|---|---|---|---|
Screenshots
The app is secured to protect government information. Screenshots of the app with fictitious data can be found below to see a sample of the app’s appearance:
Built With
- Python Flask
- Plotly Dash
- Hosted on Heroku
