SherPulse

Inspiration

Lahore loses thousands of civic complaints every day to WhatsApp groups, phone calls, and word of mouth. We kept noticing the same pattern: a pothole gets photographed and shared in a neighborhood group, gets a hundred "yes we have this too" replies, and then disappears — no tracking, no accountability, no resolution. At the same time, municipal field workers have no structured way to prioritize their day, and city administrators are making decisions without real, real-time data. We wanted to build the missing link between these three groups — citizens, field workers, and city admins — and SheherPulse was born out of that gap.

What it does

SheherPulse is a Smart City Intelligence & Civic Management Platform built specifically for Lahore, Pakistan. It brings three connected experiences into one platform:

Citizen Portal — report civic issues (waste, waterlogging, potholes, broken streetlights) with a photo that gets AI-classified by category and severity, track live incidents and air quality on an interactive GIS map, and earn Karma Points and rewards for verified reports.
Field Worker Portal — a lightweight, mobile-optimized dashboard where municipal staff receive assigned tasks on a Kanban board, get auto-optimized routes to save time and fuel, and upload before/after proof photos to verify a job is resolved.
City Admin Command Center — an executive analytics suite with live city metrics (active incidents, average resolution time, cleanliness index), priority heatmaps, AI-generated insight alerts (e.g. surges in waterlogging reports after rainfall), and ward performance leaderboards to guide workforce dispatch.

How we built it

We built SheherPulse as a modern, production-style web app using React with TypeScript, Tailwind CSS for a dark-slate glassmorphism design system, and Lucide icons for a consistent visual language. Leaflet.js and OpenStreetMap power the multi-layer GIS map, centered on Lahore with quick-zoom shortcuts to key zones like Gulberg, Johar Town, and DHA. Supabase serves as the backend, with tables designed for incidents, AQI readings, field teams, and citizen profiles, alongside authentication and real-time data sync. We pre-populated the app with realistic mock data for Lahore's wards, incidents, and AQI stations so the experience feels alive and functional from the very first load.

Challenges we ran into

Designing three genuinely different user experiences — citizen-facing, field-worker-facing, and admin-facing — inside a single cohesive app was harder than it sounds; each role needed its own information hierarchy without the product feeling fragmented. Getting the multi-layer map right was another challenge, since incidents, AQI heatmaps, and live fleet tracking all needed to render together without becoming visually overwhelming. We also had to think carefully about how to simulate "AI classification" and "route optimization" in a way that felt genuinely useful and realistic rather than gimmicky.

Accomplishments that we're proud of

We're proud of building a single platform that actually closes the loop between citizens, field teams, and city leadership — something Lahore doesn't currently have. We're especially happy with the Interactive GIS map with its layered civic-incident, AQI, and fleet-tracking views, and with the gamified citizen experience that turns a mundane complaint into something rewarding. The overall visual identity — the dark glassmorphism theme with emerald, amber, cyan, and crimson accents — also came together in a way that feels distinctly "smart city" rather than generic.

What we learned

We learned a lot about designing multi-role platforms where the same underlying data (an incident, a location, a status) needs to be presented completely differently depending on who's looking at it. We also gained a much deeper appreciation for the operational side of municipal work — things like route optimization and resolution verification aren't just UI features, they represent real inefficiencies that cost cities time and money every day. On the technical side, we learned a lot about combining GIS libraries, real-time backends, and component-driven design systems into one performant app.

What's next for SheherPulse

Next, we want to move from mock AI classification to a real trained image-classification model for incident categorization and severity scoring. We'd also like to integrate live government and WASA data feeds for AQI and drainage status, add push notifications so citizens get real updates on their reports, and pilot the Field Worker Portal with an actual municipal team in a single ward to validate the route-optimization and dispatch workflow in the real world. Longer term, we see SheherPulse expanding beyond Lahore to other major cities across Pakistan.