Polymarket Analyzer

Problem

Polymarket exposes positions and trade history on-chain but offers no first-class performance analytics comparable to traditional trading platforms. Traders wanting realised/unrealised PnL, drawdown, and behavioural heatmaps must compute them manually from raw API responses. A dedicated dashboard keyed on a wallet address or username closes that gap.

Approach

• Address/username lookup as the single entry point — no account system, no sign-in.
• Compute metrics server-side in a Bun HTTP server with TTL-based in-memory caching to absorb repeat queries without hammering the Polymarket API.
• Recharts for visualisation over heavyweight charting libraries to keep the bundle small.
• TanStack Router + Query on a React 19 frontend for typed routing and request deduplication.
• Stateless backend: cache is ephemeral, simplifying deployment and invalidation.

Implementation

Trader Analytics

Computes realised and unrealised PnL from position and fill data, plots an equity curve over time, tracks maximum drawdown, and renders a daily PnL heatmap for pattern recognition.

Position Tracking

Displays active and historical positions with per-market breakdown, win rate, and average return metrics.

Architecture

React 19 frontend with TanStack Router/Query consumes a Bun HTTP backend. Backend fetches from the Polymarket API, applies TTL-based in-memory caching, and returns normalised analytics payloads. Query deduplication on both the server cache and TanStack Query minimises redundant upstream calls.

Outcome

• Wallet-address-to-analytics flow with no account setup required.
• Equity curve, drawdown, daily PnL heatmap, and per-market win-rate breakdown.
• Lightweight Bun server with in-memory caching suitable for single-instance deployment.

Technologies

• Runtime: Bun
• Frontend: React 19, TanStack Router, TanStack Query, Recharts
• Backend: Bun HTTP server
• Data: Polymarket API, wallet address resolution