Building Accountability Through Data
FixIndia is a highly performant civic accountability platform. We use geospatial boundaries automated AI scraping and community consensus algorithms to map infrastructure failures and bring absolute transparency to local governance.
The Mission and Societal Impact
Before diving into the architecture it is critical to understand why we are building this. FixIndia bridges the gap between daily civic frustration and actionable undeniable data.
What The App Actually Does
FixIndia builds a unified and interactive map directly from the community. When a user uploads a picture of a broken road or clogged drain the app automatically maps it identifies the accountable MLA and tracks the entire process. We also scrape government portals to find exact budget sanctions and public subsidies and we link that financial data directly to the specific area or ward. This shows citizens exactly what the government is funding or failing to fund in their specific neighborhood.
How It Benefits Society
It eliminates plausible deniability. By quantifying political inaction through a Wall of Shame calculated by pure math we shift power back to the citizens. It forces authorities to prioritize repairs based on verified community density rather than political favoritism and it rewards civic engagement with actual transparency.
Data Ingestion and Verification Logic
To prevent spam and ensure the integrity of the civic map every report goes through a strict automated and community verified pipeline before it becomes public data.
1. The EXIF Metadata Pipeline
When a user uploads a picture of a civic issue the backend does not rely on manual input for the location. Instead the API strips the embedded EXIF metadata from the image. It extracts the exact GPS coordinates and the original timestamp. This guarantees authenticity and prevents users from uploading old pictures downloaded from the internet.
2. Geospatial Ward Detection
Once the coordinates are extracted the system runs an automated spatial intersection query. It does not just place a pin it identifies the accountable authority.
The backend executes a spatial point creation then runs a Point in Polygon query against our mapped wards table. It automatically enriches the report with the exact Ward Name the corresponding MLA Name and the ward Sanctioned Budget in milliseconds.
3. The Consensus Verification Machine
A newly mapped issue is never published immediately. It is governed by a strict state transition logic driven by the neighborhood
- Step 1 Submission The report enters the database locked in a pending verification state.
- Step 2 Consensus The issue becomes visible only to users currently physically located in that specific neighborhood to verify it.
- Step 3 State Change Once the database counter hits three verifications the status flips to open and it becomes visible across the city. If the community flags it as Invalid it instantly drops to resolved.
Reputation and Accountability Algorithms
The platform gamifies civic duty for citizens while strictly quantifying the performance of elected officials using purely math driven leaderboards.
The Civic Sense Score
Users build a permanent digital reputation based on their accuracy and helpfulness. The backend calculates a dynamic Civic Sense Score on the fly using a simple formula
This ensures that users who consistently verify other people reports rank higher than users who just spam complaints building a highly trusted core community.
The Wall of Shame Logic
The platform removes political bias entirely. The Wall of Shame is generated by a direct database view that aggregates the open issues. It groups them by the MLA name and sorts them. The MLA with the highest absolute count of unresolved and verified issues in their jurisdiction naturally floats to the top of the leaderboard.
The Core Architecture
FixIndia runs on a low latency and high throughput three tier architecture optimized for mobile rendering and heavy spatial database queries.
-
High Throughput API Layer
The backend runs on Bun leveraging the Elysia framework. The spatial fetching is driven by a dynamic map context endpoint. As the user pans the map the bounding box coordinates are sent to the API which streams localized reports and clusters them seamlessly.
-
AI Scraper Intelligence
To provide financial context alongside complaints unstructured government portals are scraped through an isolated ingestion loop. The system actively hunts for government budget sanctions and civic subsidies. The raw HTML is routed to a Llama AI pipeline which extracts the financial values and identifies the exact ward or area the money was assigned to. The system then executes spatial deduplication matrices checking title similarity within a thirty day window for the specified district coordinates before pushing this financial data to the map.