1

A Geographic Analysis of Homeless Management Information System

(HMIS) Data for North GA

Tim BranscombGeog 596A Capstone ProposalPenn State MGIS ProgramMay 2015

2

Outline

• Background• Goals• Project Phases • Potential Focus Areas• Project Status• Schedule • Beyond Final Project

3

Background

• Homelessness in Atlanta

• Multiple homeless organizations in Atlanta

• Desire to utilize GIS for homeless purposes

• Arrival at the Pathways Organization

• What is Pathways?

• Discussions lead to using GIS with Homeless

Management Information Systems (HMIS) data

• What is HMIS?

4

Project Goals

• Apply skills learned in MGIS program to aid homelessness

• Provide customer focused geographic analysis

• Create template of GIS procedures for other HMIS providers to follow

• Create ‘path’ for data sharing between Pathways and other requesting organizations

5

Project PhasesPhase 1-Data Phase 2-Analysis Phase 3-Presentation

6

Phase 1 – Data

• Research HMIS data

• Investigate available census data estimates and entities

• Locate tools for pre-processing (Excel, Access, ArcMap)

• Execute required training and agreements

• Establish accounts needed for data access

7Data Dictionary Data Model

Data Standards Manual

HMIS: Homeless Management Information System

Phase 1 – Data

8

• Basic review of Census Bureau geographic entities, their relationships, and available datasets

• Non-alignment between zip code boundaries and traditional census entities (tracts, block groups, etc.)

• Close alignment between zip code boundaries and Zip Code Tabulation Areas (ZCTAs)

U.S. Census Bureau Information

Phase 1 – Data

9

Use Iterative workflow to answer general questions:

• What areas have the highest concentration of clients?

• What types of clients are from which areas?

• What types of service do the clients use the most?

• What distances do clients have to travel for services?

• Which areas have an increasing/decreasing rate of clients?

Phase 2 – Analysis

10

Original HMIS Report Data

Excel Pivot Table to Summarize Data

Imported and Linked to ZCTAs

Resulting Geographic Visualization(s)

Phase 2 – AnalysisData Pre-Processing Steps

11

Phase 2 - Analysis

Thematic Map of Clients with Children Rates per Zip Code

Proportional Symbol Map of Client Totals per Zip Code

12

Phase 2 - Analysis

Pie Chart Symbol Map of Client Race Proportions per Zip Code

Drive-Time Polygon Map from Family Focused Shelters

13

Phase 2 - Analysis

Hotspot Analysis of Veteran Rate per Zip Code

Grouping Analysis Map of Homeless Client Rates per Zip Code

14

• Most ‘telling’ results are selected by Pathways

• Storyboard created by Pathways for the selected results

• Publish appropriate feature services to support storyboard

• Story Map created and published after iterative process

• Appendix documentation created for selected data and maps

Phase 3 - Presentation

15

• Initial Inquiry

• Outcome (map)

• HMIS entities used

• Required SQL

• Census data used

• Pre-processing steps

• GIS workflow

Report Documentation

Phase 3 - Presentation

16

• Regression modelling

• Custom modification of Story Maps (using JavaScript)

• Significant focus on template for other HMIS organizations

• Windows utility application creation for pulling and aggregating data

as needed (via ODBC connection)

• ArcMap Python scripts and/or models for pre-processing needs

• Excel VBA scripts for converting data and/or producing necessary

pivot tables

Potential Focus Areas

17

Project Status

18

February – May 2015 Phase 1 – Data Research and Coordination

April – July 2015 Phase 2 – Analysis

June – September 2015Phase 3 – Presentation (Templates and Story Maps )

September 2015Draft final paper and conference presentation

October 2015 Present Results at National HMIS Users Conference (Washington, DC)

Project Schedule

19

• Continued involvement/analysis for Pathways

• Obtain non-profit ArcGIS software for Pathways

• GIS Training for Pathways

• Take basic approach to one of several non-profit

organizations I would like to work GIS for

• Adapt methods to open source products

Beyond Capstone Project

2016…

Ref: ESRI 2015

20

Dr. Douglas Miller – Project Advisor

Dr. Josie Parker – Pathway’s Research Project Manager

Dr. Jack Barile– Pathway’s Data Researcher

Dr. Justine Blanford – Future geo-statistical consultant

Meghan Branscomb – Supporting Wife!

Acknowledgements

21

Census.gov (2015a). Zip code Tabulation Areas (ZCTAs). Retrieved April 2, 2015 from https://www.census.gov/geo/reference/zctas.html Census.gov (2015b). American Community Survey: When to use 1-year, 3-year, or 5-year estimates. Retrieved April 2, 2015 from http://www.census.gov/acs/www/guidance_for_data_users/estimates/ Department of Housing and Urban Development (2005). Making the Most of HMIS Data: A Guide to Understanding Homelessness and Improving Programs in Your Community. Retrieved March 20, 2015 from https://www.hudexchange.info/resource/1316/guide-to-understanding-homelessness-and-improving-programs/ HudExchange.info (2014a). Homeless Management Information System. Retrieved March 20, 2015 from https://www.hudexchange.info/hmis/ HudExchange.info (2014b). HMIS Data Dictionary. Retrieved March 20, 2015 from https://www.hudexchange.info/resource/3824/hmis-data-dictionary/ Loubert, Linda (2010). Mapping Urban Inequalities with GIS. Retrieved March 20, 2015, from http://www.esri.com/news/arcnews/spring10articles/mapping-urban.html Olivia, Jon-Paul (2006). Using Geographic Information Systems (GIS) as a tool for HMIS decision making. Retrieved March 20, 2015 from https://www.hudexchange.info/resource/1572/using-gis-as-a-tool-for-hmis-decision-making/ PCNI.org (n.d.) Pathways Community Network Institute. Retrieved March 20, 2015, from http://www.pcni.org/about-us Storymaps.argis.com (n.d.) Use StoryMaps to Inform and Inspire Your Audience. Retrieved March 20, 2015, from http://storymaps.arcgis.com/en/ Wong, Yin-Ling I, Hiller, Amy E. (2001). Evaluating a Community Based Homelessness Prevention Program: A Geographic Information System Approach. Administration in Social Work 25:4, pp21-45.

References