AI-assisted spatial equity research for community development organizations.
Community Data Lab is an open-source methodology, curriculum, and toolkit that teaches nonprofits to answer causal questions about their neighborhoods using AI-assisted workflows and public data.
Built by TresPies. Proven at Commonwealth Development in Madison, WI.
Community development corporations (CDCs) need spatial equity research to advocate effectively — but the existing tools solve the wrong problems:
| Tool | What it solves | What it doesn't |
|---|---|---|
| Esri Nonprofit Program | Software access ($100/yr) | Research methodology |
| QGIS | Cost (free) | Data sourcing, analysis design |
| Urban Institute Spatial Equity Tool | Quick disparity snapshots | Custom research questions, causal analysis |
| PolicyMap | Ready-made neighborhood data | Original research, policy brief production |
| GIS consultants | Maps and analysis | Capacity building — they leave, the knowledge leaves |
The gap: Nobody teaches CDCs to design and execute original spatial equity research. CDCs with GIS licenses produce maps that illustrate rather than argue.
A reproducible research workflow for nonprofits:
- Research Question Framing — What makes a spatial equity question answerable with public data
- Data Sourcing — Census API, GTFS, state education data, HRSA, HUD, Eviction Lab — all public, all free
- AI-Assisted Analysis — Using AI as a research partner (not a chatbot) for regression, threshold detection, outlier identification
- Output Production — Standalone HTML data viewers, policy briefs, funder-ready narratives
- Maintenance — Refresh schedules, data quality monitoring, institutional memory
A structured training program (lesson plans and exercises coming Q3 2026):
- Week 1: Spatial equity questions — framing what's answerable
- Week 2: Public data landscape — Census, transit, education, health, housing sources by state
- Week 3: AI as research partner — prompt engineering for spatial analysis
- Week 4: Analysis patterns — OLS regression, composite scoring, outlier detection (stdlib Python, no dependencies)
- Week 5: Output production — standalone viewers, policy briefs, Excel workbooks
- Week 6: Maintenance and sustainability — refresh cycles, coalition calendars, institutional memory
Parameterized Python scripts (stdlib only, no pandas/numpy) that can be configured for any US city:
- Census API data fetchers (demographics, income, housing cost burden, uninsured rates)
- GTFS transit frequency analyzer
- Composite equity scoring engine
- Standalone HTML viewer builder (Leaflet.js, embedded data, no server required)
- Policy brief templates (transit, housing, education, zoning, rapid response)
AI prompt templates for each stage of the research workflow. Designed for Claude Code, adaptable to other AI tools.
Commonwealth Development (Madison, WI) used this methodology to produce:
- 7 causal research question analyses answering "does X predict chronic school absence after controlling for income?" across transit, housing, health, and childcare domains
- 11-layer interactive equity data viewer (3.8 MB standalone HTML, no server, PDF export)
- 19 verified public data sources with automated refresh pipelines
- 2 quasi-experimental research designs submitted to Arnold Ventures
- 5 policy brief templates for city council and funder audiences
- Agent memory architecture maintaining institutional research knowledge across sessions
All built with Claude Code, stdlib Python, and zero external infrastructure.
community-data-lab/
curriculum/ # 6-week cohort materials
methodology/ # The TresPies research workflow documentation
prompt-library/ # AI prompt templates for each research stage
starter-kit/ # Parameterized Python scripts and HTML templates
case-studies/ # Partner city analyses and lessons learned
templates/ # Policy brief, viewer, and workbook templates
CDC Research Accelerator is TresPies' consulting practice built on Community Data Lab methodology. When you hire TresPies, you get:
- City-specific data infrastructure built and delivered
- Custom standalone viewers for your portfolio and service area
- Policy briefs grounded in your spatial equity data
- Staff training using Community Data Lab curriculum
- Ongoing research partnership
Contact: Cruz Morales, TresPies — cruz@trespies.com
Community Data Lab is open source. Contributions welcome:
- New data source adapters — State-specific education, health, or housing data parsers
- City starter kits — Pre-configured scripts for specific metro areas
- Curriculum translations — Bilingual materials for specific communities
- Case studies — Document your organization's experience with the methodology
MIT License. Use it, fork it, teach with it.
Built by TresPies. Proven at Commonwealth Development.