Weather Splits Are Now on FanGraphs!

Weather data is now available on the FanGraphs Splits Leaderboards.

You can filter any data since 2010 by temperature, barometric pressure, elevation, wind speed, wind direction, and air density. To make the weather data more useful, you can also filter on the specific ballpark, roof type, and the roof open/closed status.

The new filtering options include:

• Ballpark Type: This includes outdoor, fixed dome, and retractable roof filters, as well as separate game status filters for roof open, roof closed, and dome.
• Weather: This allows you to filter on general weather conditions such as rain, fog, clouds, and haze.
• Temperature Range: Select a temperature range in Fahrenheit.
• Barometric Pressure: Select a barometric pressure range.
• Air Density: Select an air density range. We are using the calculation available here, which is adjusted for altitude, temperature, humidity, and pressure. In general, it will correlate strongly with temperature.
• Wind Direction: Choose the direction the wind is blowing. Blowing to LF, Blowing to CF, and Blowing to RF are in 30 degree segments, while Blowing Out is the full 90 degrees towards the outfield. Blowing In is the full 90 degrees towards the infield. Blowing from 1B or 3B is the full 90 degree crosswind.
• Wind Speed – Filter on the current wind speed in MPH.

Weather conditions can change drastically (up to 30 degree) during a game, which in theory can make a big difference when doing weather-related baseball research. So instead of taking the starting weather conditions and applying them to the entirety of the game, our weather data (provided by OpenWeather) is available hourly; we’ve matched the timestamp of each plate appearance to the nearest hour.

For the most part, you’ll find the weather data results intuitive. For instance, you will generally have more offense in warmer weather in open air parks:

The same goes for lower air density in open air parks:

And if you thought wind impacted offense for fly balls or line drives, you’d be correct:

With the splits leaderboards, you’ll be able to look up all the weather data you like and do your own research. Still, keep in mind that this is only a very broad look at weather. This data is not normalized for parks or other factors because it’s less necessary due to the incredibly large sample.

Have fun, and since this is a little bit of an experiment for us, please let us know if you have any suggestions for how to make our weather data more useful.