Skip to content

jscarlton/nfldata

 
 

Repository files navigation

nfldata

The are two series of MapReduce programs. One is a series of programs to extract and normalize the data. The second is a simple program to look at incomplete passes.

The play by play dataaset can be found at http://www.advancednflstats.com/2010/04/play-by-play-data.html.

ETL Series

This program takes the play by play dataset and merges it with other datasets like arrests, stadiums and weather.

Set things up by running the setup.sh script or run the following steps manually:
Run the PlayByPlayDriver on the play by data data.
Run the ArrestJoinDriver on the data from PlayByPlayDriver. (place in HDFS under joinedoutput)
Put the stadiums.csv in HDFS in a directory called stadium.
Put the 173328.csv in HDFS in a directory called weather.
In Hive, run playbyplay_tablecreate.hql.
In Hive, run playbyplay_join.hql.
In Hive, run adddrives.hql.
In Hive, run adddriveresult.hql.
Query and have fun!
See the queries in queries.hql for some examples of how and what to query.

Incomplete Passes

Simple MapReduce on NFL play by play data. This program focuses on incomplete passes and which receiver they were throw to. See http://www.jesse-anderson.com/2013/01/nfl-play-by-play-analysis/ for the resulting charts.

About

Combining datasets with MapReduce on NFL play by play data.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Java 91.2%
  • PigLatin 4.1%
  • Python 3.1%
  • Shell 1.6%