In 2023, the City of San Diego purchased dozens of “ice fishing tents” as emergency shelter for unhoused residents.
The assumption?
“Ice fishing” meant insulated and warm.
The reality?
These tents were designed for use with an external stove — and included a vent hole on top and no floor.
The result?
Leaks, cold air, rodents, and frustration from the very people the program aimed to help.
This isn’t a failure of intent — it’s a failure of feedback.
Without real-time insights and lived-experience context, even well-funded programs can miss the mark.

Combines point-in-time counts, live reporting, and real-time kiosk feedback
Surfaces gaps between what’s deployed and what’s actually needed
Supports smarter, more empathetic decision-making — backed by data, not guesswork


Verified homeless population data
Reported concerns from Get It Done
Transit accessibility (via SANDAG)
Rent burden and demographic pressure zones
Existing service locations (via 211 SD)


GIS overlays of point-in-time homelessness counts
Live Get It Done reports processed with OXEN.AI (turning messy complaint text into structured data)
Demographic risk indicators like rent burden and age
Transit overlays from SANDAG
Recommendations for what type of service (e.g. mobile clinic, hygiene station, outreach) to deploy — and where
A simple map-based interface to find nearby shelters, showers, food, healthcare, and legal services
Transit routes and walkable options
Real-time updates when resources change
ADA-friendly design for maximum accessibility





Every time someone searches for a shelter, looks up a mobile clinic, or checks transit options on a Vulnerability Atlas kiosk, that activity is logged (anonymously) and fed back into the Service Deployment Engine.
Over time, this helps identify:
High-demand services that may be missing in certain regions
Mismatch between availability and need (e.g., areas where people look for showers but none exist nearby)
Emerging patterns based on time of day, weather, or event-driven spikes
📊 It’s not just a tool for users — it’s a feedback signal for the system.
By combining public reports, official datasets, and real-time kiosk usage trends, Vulnerability Atlas continuously evolves — making smarter, more empathetic recommendations as it learns.

















