Nigran Lahore

Watch Over Lahore.

Comment

Inspiration

Lahore already generates a huge amount of urban information—weather observations, air-quality measurements, geographic data, environmental studies and operational records. The challenge is that this information is usually experienced as separate numbers, dashboards or reports.

We wanted to explore a simpler question:

What does all of this data mean for a specific place, right now—and what should happen next?

That idea led to Nigran Lahore, a hyperlocal urban intelligence platform designed to turn fragmented city signals into understandable risk information and actionable decisions for citizens and city teams.

Rather than building another dashboard that simply displays data, we wanted to build a system that helps users see the risk, understand the reason, and act early.

What it does

Nigran Lahore combines environmental and urban signals into a single, explainable risk view.

For citizens, Nigran provides a location-based view of:

Flood risk Air-quality risk Heat risk Waste/drain vulnerability

It combines these signals into a hyperlocal risk picture and provides a six-hour outlook, plain-language explanations, and recommended actions.

For city-operation scenarios, Nigran provides a prototype priority queue that ranks locations according to factors such as environmental conditions, simulated drainage/waste state, service age and population exposure.

The key feature is “Why?”. Instead of showing only a score, Nigran explains the factors contributing to that score so users can understand how the recommendation was produced.

The prototype currently uses live weather and air-quality data, publicly available/reference geographic and environmental information, and explicitly labeled simulated operational telemetry where real-time municipal sensor data is not publicly available.

How we built it

Nigran Lahore is a software-only web application designed around free and accessible data sources.

Technology React + Vite Tailwind CSS Leaflet / OpenStreetMap TypeScript Open-Meteo weather and air-quality APIs Browser local storage for lightweight state/cache Explainable rule-based risk engine Responsive web interface Vercel deployment

The risk engine combines multiple inputs into bounded 0–100 scores. Rather than pretending to have a trained ML model where sufficient labeled Lahore data is unavailable, we use an explainable weighted scoring framework whose contributing factors can be inspected directly.

The application also includes a data-status system that distinguishes live, stale and snapshot states, so an API failure does not silently appear as a safe environmental condition. The audit verified the production weather and air-quality values against the raw API responses and found exact matches.

Challenges we ran into

The biggest challenge was data availability versus real-world ambition.

Lahore has many relevant urban systems, but not every operational dataset is publicly accessible as a live API. We therefore had to design Nigran around what could actually be verified remotely.

That led us to three principles:

Use real data where it is available. Weather and air-quality inputs are retrieved from live sources.

Use public/reference data transparently. Historical and geographic information is treated as reference material rather than presented as live telemetry. Simulate only what must be simulated for the prototype. Drain and operational telemetry is explicitly treated as simulated prototype data rather than being presented as real municipal sensor information. We also had to balance ambition with hackathon time. Several planned features, such as full safe-route calculation and nearest-resource routing, were deliberately cut so that the core risk and decision-support experience remained reliable.

Accomplishments that we're proud of

We are especially proud that Nigran is not just a visual concept—it has a functioning decision layer behind the interface. Highlights Live Lahore weather data verified value-for-value against the raw API Live air-quality values verified against the raw API Six-hour forecast logic validated Explainable risk scores bounded between 0–100 Graceful handling of missing weather data Priority queue dynamically responds to changing conditions Citizen and field-operations views implemented 77/77 unit and component tests passed in the audit Production build and deployment verified No exposed API keys or localhost references Clear distinction between live and simulated prototype data The feature we are most proud of is probably the combination of risk score + explanation + action. A user doesn't just see “78”—they can understand why the score is high and what the system recommends doing next.

What we learned

We learned that building a smart-city product is less about collecting the most data and more about making data useful at the point of decision. We also learned that: A live API is valuable only when its data is correctly interpreted and timestamped. Missing data must never silently become a “safe” condition. Explainability is important when a system produces risk scores. A prototype can still be useful without pretending to have access to proprietary infrastructure. Scope control is critical: a smaller working decision system is stronger than a larger collection of unfinished features. Good data provenance and clear labeling are part of the product—not just documentation. Most importantly, we learned to distinguish what the city could eventually support from what we can honestly demonstrate today.

What's next for Nigran Lahore

Nigran Lahore is currently a prototype, but the architecture is designed to grow beyond the hackathon. Future versions could connect to additional verified Lahore data sources, including richer historical flood information, municipal waste data, drainage information, environmental sensors and other open-data feeds. The long-term goal is to move from: “Here is the risk.” to: “Here is the risk, here is why it is happening, here is who is affected, and here is the best next action.” We also want to improve the predictive layer using larger historical datasets, validate the risk models against real events, and eventually provide more advanced operational planning tools for city agencies. Nigran Lahore is intended to become an intelligence layer that cities can build on—not another dashboard that only displays what is already known.

Built With

Share this project:

Updates