Kalshi Trading Personality

Problem

Kalshi traders accumulate extensive position history but have no native tool that translates that history into a behavioural profile or surfaces markets aligned with their revealed preferences. Generic analytics dashboards report PnL without characterising how a trader trades. A lightweight application that classifies trading style and recommends markets on that basis fills this gap.

Approach

• Kalshi API as single data source: Pull trading history directly rather than scraping or relying on third-party aggregators.
• Personality classification over raw metrics: Map positions and trade cadence to discrete trader archetypes for shareability and recall.
• Hono + Bun backend with SQLite for a minimal, low-latency server footprint; no external database dependency.
• TanStack Router/Query on React for type-safe client routing and server-state caching.
• shadcn/ui for consistent, accessible components without a heavyweight design system.

Implementation

Trading Analysis

Authenticates against the Kalshi API and ingests a user’s historical positions and fills. Derives a trading personality type from activity patterns and produces personalised market recommendations scoped to the inferred preferences.

Shareable Profiles

Auto-generates trader profile cards summarising the classified personality and recommended markets. Each profile is exposed at a shareable link for social distribution.

Architecture

Hono backend persists analysis results in SQLite. React frontend consumes backend routes via TanStack Query, with routing handled by TanStack Router and UI built on shadcn/ui primitives.

Outcome

• End-to-end flow from Kalshi API credentials to a shareable personality profile.
• Personalised market recommendations grounded in historical trade data.
• Self-contained stack (Bun runtime, SQLite file) deployable without managed services.

Technologies

• Backend: Hono, Bun, SQLite
• Frontend: React, TanStack Router/Query, shadcn/ui
• API: Kalshi Demo API