The Retrosheet database is a well-known and largely community effort to digitize play-by-play information for both historical and modern major league baseball games.

This visualization shows schedules, locations, and weather conditions for baseball games recorded in the Retrosheet database since 2001. Views include tables of stadiums, home teams, and away teams (with nested time series of game occurrences for each one), a wrapping multi-calendar of game schedules (using our reruns technique), a geographic map of stadium locations, a scatterplot matrix of dates and three other dynamically-assignable data dimensions, and a drill-down tables of detailed game information.

The visualization uses the cross-filtered views technique to support high-dimensional brushing between the various views. For example, in the screeshot above the game calendar view, map, games drill-down table, and stadium condition variations scatterplot matrix are filtered to show only games involving selected away teams (Atlanta Braves).

The key interaction in the visualization is variation filtering, which can be toggled on and off using the checkbox near the bottom right corner. When on, dragging and stretching the rubberbands in the variations scatter plots dynamically filters all other views to show only games within the three rubberband ranges (i.e. a 3-D cubic region of stadium condition values). This makes it possible to observe the effect of temperature, wind speed, attendance, etc. on geographic and temporal game distributions interactively.

Visualization (6.1 MB)