Dota 2 drafting involves two team captains taking turns banning and picking heroes in sequence. The Dota 2 web API allows for access to the almost 2 billion Dota 2 games. The goal will be to have a model for suggesting picks and bans, trained on the existing Dota 2 game records.
Use: httprequest.py grabs data from Valve servers and writes it as a csv. Send the csv through MapReduce with the mapper and reducer py files, and parse the output. The result can then be graphed.
A simple pandas script that concatenates all the data csv files in "\collection" into a single output file "capmodedata.csv" for later analysis.
A pandas script currently used to graph the most popular picked/banned heroes. Uses the output from mapreduceoutputparse.py, which itself depends on successful execution of MapReduce on the data set.
A simple lookup for hero names based on id values.
JSON object containing hero id names and localization tags.
httprequest.py is a python script that combs the Dota 2 web API for captain's mode games and stores the records in "\collection\data.csv" for later analysis. It requires two files, not included:
- key.dat - a file that has only the API key in it (http://steamcommunity.com/dev/apikey)
- record.dat - a file with only the "seq num" that the program will start collecting at. Consult the Dota 2 web API documentation for which seq num you wish to start at.
httprequest.py will start at the seq num in record.dat and grab matches until a fixed seq num, 1704492188. Edit httprequest.py for more matches as desired.
Components of a MapReduce framework. Invoke like hadoop jar contrib/streaming/hadoop-*streaming*.jar -file /dir/mapper.py -mapper /dir/mapper.py -file /dir/reducer.py -reducer /dir/reducer.py -input /dir/* -output /dir/dir-of-output
Loads the MapReduce output into a pandas dataframe for further processing.