Game Insights

Game Insights is a local-first analytics app for game teams. It imports tabular data, detects the game type, runs AI-assisted analysis, and surfaces dashboards for upload, funnels, monetization, games, and settings.

What It Does

• Imports CSV, TSV, JSON, NDJSON, Excel, and SQLite data
• Supports file upload, URL import, clipboard import, and folder import
• Auto-detects game type and maps columns with AI-assisted analysis
• Persists data locally in the browser
• Surfaces specialized views for funnels, monetization, AI analytics, games, and dashboard building

Key Screens

• Upload for importing and reviewing data
• AI Analytics for insights, anomalies, and Q&A
• Funnels for detected or manual funnel analysis
• Monetization for revenue and payer analysis
• Dashboard Builder for custom dashboards
• Games for multi-game management
• Settings for theme, language, and AI provider configuration
• Landing for the public marketing page

Repository Structure

src/
├── ai/              # App analytics engine: schema, metrics, funnels, cohorts, recommendations
├── adapters/        # Data source adapters
├── components/      # Shared UI, upload flow, charts, landing, settings
├── context/         # React state providers
├── hooks/           # Shared hooks for analytics, onboarding, keyboard shortcuts
├── lib/             # Importers, stores, persistence, utilities, sample data
├── pages/           # Route-level screens
├── services/ai/     # LLM provider orchestration, prompts, chains, tools, memory
├── services/openai.ts
└── types/           # Shared TypeScript types

The two AI areas are intentionally separate:

• src/ai contains domain analytics used by the app itself
• src/services/ai contains provider-backed LLM orchestration

Quick Start

Prerequisites

• Node.js 18+
• pnpm

Install and run

pnpm install
pnpm dev

Open http://localhost:5173.

Sample data

On the upload page, use Try Example Data to load one of the built-in datasets:

• Puzzle Game Analytics
• Idle Game Analytics
• Gacha RPG Analytics

Commands

pnpm dev             # Start the Vite dev server
pnpm build           # TypeScript check + production build
pnpm lint            # ESLint
pnpm test            # Vitest in watch mode
pnpm test:run        # Vitest once, then cleanup
pnpm test:coverage   # Vitest with coverage
pnpm test:e2e        # Playwright end-to-end tests
pnpm storybook       # Storybook
pnpm build-storybook # Build Storybook static output

Architecture

The current app flow is:

Upload / sample data -> importer -> local storage -> AI analysis -> dashboard screens

Core modules involved in that flow:

• src/lib/importers/ parses supported file and text sources
• src/lib/dataStore.ts and src/lib/db.ts handle local persistence
• src/lib/columnAnalyzer.ts bridges optional AI-backed column analysis
• src/ai/ computes metrics, funnels, cohorts, recommendations, and report outputs
• src/services/ai/ wraps provider configuration, prompts, chains, tools, and memory
• src/context/ holds app-wide state for data, games, theme, and toasts

Design Notes

• The app uses CSS variables for the warm theme in src/index.css
• Most dashboards and flows are code-split with React.lazy
• Shared UI lives in src/components/ui/

Related Docs

• Architecture
• Usage Guide