LottoPickle is a data science project built on top of New York State's publicly available lottery dataset. The goal: demonstrate that a single data source, interrogated from multiple analytical angles, can produce genuinely useful and visually compelling insights.
The premise is simple: take a publicly available dataset with well-understood properties (random draws, fixed number ranges, long history) and apply every major data science technique to it — frequency analysis, time series, inferential statistics, combinatorics, probabilistic modeling.
The lottery framing makes the analysis accessible and fun. But the underlying methods — heatmap encoding, cumulative trend curves, co-occurrence hash maps, skewness coefficients — are the same techniques used in finance, genomics, retail analytics, and recommendation systems.
All computation runs client-side in the browser using React's useMemo hooks, eliminating the need for a backend entirely. Data is fetched live from the New York State SODA API — up to 5,000 draw records per game, parsed and aggregated on the fly.
The result is a fully functional, zero-latency analytics dashboard that demonstrates practical application of data science concepts in a production web environment.
Built from scratch as a data science platform using publicly available New York State lottery data.
Frequency analysis, heatmaps, trend curves, distribution statistics, co-occurrence pairs, gap analysis, and probabilistic number generation.
Active development. New analytical views and data sources in progress.
All analysis uses historical Powerball and Mega Millions draw data from the New York State government open data portal (data.ny.gov). The dataset is publicly available, free to access, and updated after every draw.
Data is fetched via the SODA REST API, requesting up to 5,000 records ordered chronologically. No data is stored — everything is computed fresh on each session in the user's browser.
Read the full methodology →