We monitor 100s of live data streams across transport, energy, air quality, satellite imagery, and financial markets, discovering cross-domain patterns no single agency can see.
London’s collective mood predicts grid demand 6 hours ahead, r = 0.92
Fox population decline + new cafe filings predict gentrification 6–12 months early
Social media surge + traffic anomalies predicted protest location and size 48 hrs ahead
Data
Transport for London
National Grid ESO
Environment Agency
Met Office
UKHSA
UK Parliament
Thames Water
Metropolitan Police
HM Land Registry
UK Power Networks
Port of London Authority
London Air Quality Network
Breathe London
100+ sources and counting
Discoveries
Real findings discovered autonomously by cross-referencing London's data streams.
The city’s collective social sentiment is a statistically significant leading indicator of National Grid electricity demand. When Londoners feel good, they consume more energy: shared cultural moments, mass kettle-boiling, grid spikes.
Fused Companies House filings (new cafes) with iNaturalist ecological data: fewer urban foxes, more ornamental birds. Predicts neighbourhood transformation 6–12 months before real estate models catch on.
Cross-referencing social media sentiment spikes with real-time traffic anomalies to predict where the next protest will emerge and its likely size, hours before crowds physically gather.
TfL sensors show Canary Wharf at 9% capacity, Bank at 37%, while railways restore pre-COVID Friday frequencies. The system flagged a strategic capital mismatch: billions invested in capacity nobody uses, declaring WFH Friday a permanent feature of economic geography.
National Grid shifts to gas caused Polymarket Net Zero contracts to drop within hours. Uncovered a real-time shadow referendum on climate policy, priced minute-by-minute from energy data.
Computer vision models count cyclist helmet-wearing rates across boroughs via TfL CCTV feeds. Higher helmet rates strongly correlate with higher property values and lower crime, turning bike safety into an accidental socioeconomic indicator.
How it works
Hundreds of data streams flow through a network of specialised AI agents that cross-reference domains and surface patterns no single source could reveal.
Inspired by the human eye, the system allocates processing power dynamically. Areas with clustering anomalies become the fovea, receiving dense, high-frequency analysis. The rest of London stays in peripheral vision across 4,000+ grid cells.
We're looking for partners in government, research, and infrastructure.
Daily briefing
The most interesting cross-domain patterns, anomalies, and discoveries. Delivered to your inbox every morning.
No spam. Unsubscribe anytime.