Market Prediction

altcoin_correlation_heatmap
altcoin_momentum_heatmap
altcoin_perf_heatmap
btc_dom_altseason_chart
btc_momentum_chart
sentiment_gap_bar
sentiment_gap_table
sentiment_vs_tech_scatter

Inspiration Trading crypto felt like guessing between two conflicting signals — prediction markets and technical indicators — with no way to measure which was right. We wanted to quantify that gap.

What it does Pulls live data from Kalshi, CoinGecko, AlphaVantage, and FRED, then ranks divergences between crowd-implied probabilities and technical momentum signals across 29 cryptocurrencies — outputting 8 charts in under 2 minutes.

How we built it A 7-phase Python pipeline orchestrated by a single MAIN.py, vibe-coded on the Zerve platform and polished on a personal server after token expiration.

Challenges we ran into CoinGecko's free-tier rate limits hitting mid-pipeline, Kalshi returning hundreds of unpriced initialized markets, and prediction market data only covering BTC and ETH — forcing us to bring in CoinGecko community sentiment for the remaining 27 coins.

Accomplishments that we're proud of A fully automated, end-to-end system that produces actionable, ranked divergence signals from five real data sources — with zero hardcoded credentials and readable CSV/JSON outputs alongside every chart.

What we learned Prediction market implied probability and technical momentum measure fundamentally different things — combining them surfaces edges neither source reveals alone. Also: always build retry logic before hitting free-tier APIs.

What's next for Market Prediction Expanding to on-chain data (wallet flows, exchange reserves), adding a live dashboard, and backtesting whether historical sentiment–reality gaps were genuinely predictive of subsequent price moves.