canairy preparedness
family warning system — 48hr advance notice when institutions start moving
problem
families get financial warnings after institutions, not before. During March 2023 banking crisis, ordinary people learned from headlines while hedge funds moved days earlier.
solution
22 real indicators from 15+ APIs (FRED, Treasury, FDA, Reuters) → weighted scoring → 48hr advance warning → $47 avg prep cost.
Canairy Dashboard: Real-time threat monitoring and preparedness actions.
Detailed Indicators: Tracking 22 critical signals across multiple categories.
architecture
indicator categories:
- financial: treasury stress, VIX volatility, credit spreads
- supply chain: baltic dry index, hormuz premium, port congestion
- energy: natural gas, power grid, strategic reserves
- geopolitical: taiwan strait, cyber attacks, civil unrest
- health: FDA shortages, hospital capacity
- emerging: AGI development, labor displacement
design & process reflection
the development of canairy evolved from a simple portfolio demonstration into a genuinely functional early warning system. this transformation revealed several layers of complexity i hadn't initially anticipated.
core challenges
authenticity & data reliability
evolving from mock to live data (40+ collectors) presented the challenge of maintaining reliability with external apis. my solution focused on graceful degradation and transparent data source labeling.
multi-domain integration
threading together disparate data sources (real-time to monthly updates, json/csv/pdf formats) into cohesive threat assessment required building domain-specific collection strategies.
crisis information ux
balanced "calm preparedness" with "urgent awareness" through threat-responsive branding and mobile-first design. prioritized actionable family guidance over technical metrics.
key decisions
data architecture
opted for json files over postgresql for simplicity/portability, and http polling with caching over websockets for reliability. chose monolithic flask app for easier deployment over microservices.
user interface
selected dark theme with threat-responsive elements over traditional dashboards. prioritized family-friendly preparedness actions over government-style technical alerts.
data sources
chose free/government apis over paid financial feeds for accessibility. implemented news api over social media scraping for reliability.
key learnings
multi-source data reliability: learned the critical importance of building systems that gracefully handle messy, real-world data (api failures, rate limits) to maintain user trust. a hybrid live + fallback approach proved essential.
technical sophistication vs. family usability: families need "should i go to the store today?" not "basis points." the priority actions panel with phone numbers and specific steps bridged this gap successfully.
threat visualization psychology: the canary branding communicates the concept non-threateningly. using warm amber for caution rather than aggressive reds proved crucial for user engagement.
mobile-first crisis design: during emergencies, people use phones. responsive design became core to the system's utility, not an afterthought.
final insight
sometimes the best portfolio projects are the ones that accidentally become useful. the technical challenge of building a realistic demo led to building something genuinely functional, which taught me more about full-stack development, api integration, and user experience than any purely academic exercise could have.
proof: March 2023 banking crisis alert (48hr early), live dashboard, 47 families using