Hourly weather data for the New York City Taxi Trip Duration Challange

Here is some detailed weather data for the New York City Taxi Trip Duration Challange. I noticed that many contenders use daily weather data and thought that the ML could be improved with hourly data for NYC (default KNYC station) since pickup_datetime is given. python code on github can return same data for any city

Content

Wundergrounds API provides hourly weather data in JSON format, but I assume most people just want the complete data set in csv. i stands for imperial, m for metric so the difference stands in the relative unit for the returned value (ex. Fahrenheit vs. Celsius).

Note that values will = -9999 or -999 for Null or Non applicable (NA) variables. (replaced with NaN in Version 2)
Wundergrounds full Phrase Glossary

• datetime: Date and time of day (EST)

• tempm: Temperature in Celcius

• tempi: Temperature in Fahrenheit

• dewptm: Dewpoint in Celcius

• dewpti: Dewpoint in Fahrenheit

• hum: Humidity %

• wspdm: Wind speed in kph

• wspdi: Wind speed in mph

• wgustm: Wind gust in kph

• wgusti: Wind gust in mph

• wdird: Wind direction in degrees

• wdire: Wind direction description

• vism: Vivibility in Km

• visi: Visibility in miles

• pressurem: Pressure in mBar

• pressurei: Pressure in inHg

• windchillm: Wind chill in Celcius

• windchilli: Wind chill in Fahrenheit

• heatindexm: Heat index Celcius

• heatindexi: Heat index Fahrenheit

• precipm: Precipitation in mm

• precipi: Precipitation in inches

• conds: Conditions: See full list of conditions

• icon

• fog: Boolean

• rain: Boolean

• snow: Boolean

• hail: Boolean

• thunder: Boolean

• tornado: Boolean

Thanks to Wunderground