Research Data

Deep historical archives and purpose-built training datasets for quantitative research, backtesting strategies, and machine learning on prediction markets.

Historical Data

Trade History

Full trade-level data across Polymarket and Kalshi — timestamps, prices, volumes, wallet addresses, and direction.

Orderbook Snapshots

Depth-of-book captures at configurable intervals for liquidity analysis and market microstructure research.

Market Lifecycle

Complete market metadata from creation through resolution — outcomes, settlement, open/close dates.

OHLCV Candlesticks

Price and volume data bucketed at 1m, 5m, 15m, 1h, 4h, and 1d resolutions.

Social Signals

Historical social media mentions and sentiment scores tied to prediction market events.

Wallet Profiles

Point-in-time snapshots of wallet analytics, H-Scores, and performance metrics.

Custom Training Data

Purpose-built datasets for training ML models and fine-tuning LLMs on prediction market intelligence.

Trader Behavior Patterns

Labeled datasets of trading strategies, risk profiles, and performance outcomes across thousands of wallets.

Market Resolution Signals

Feature sets correlated with correct predictions — for training forecasting and signal detection models.

Anomaly Detection Sets

Labeled price jumps, wash trading patterns, and unusual activity for classifier training.

RAG-Ready Knowledge Bases

Pre-chunked, structured market data optimized for retrieval-augmented generation workflows.

Delivery Formats

JSON / JSONL

Recommended

Structured records for APIs, data agents, and LLM ingestion.

CSV / Parquet

Tabular format for Pandas, Spark, BigQuery, and analytical tools.

API Access

Paginated endpoints for programmatic bulk retrieval on demand.

Get Access

Research data packages are tailored to your specific use case. Reach out with your requirements and we'll put together a dataset that fits.

contact@heisenberg.so