View Code? Open in Web Editor
NEW
A predictive engine to forecast the NBA All-Star team
HTML 96.65%
Python 0.65%
Shell 0.05%
Jupyter Notebook 2.66%
social-media-or-ball-game's Introduction
Ball Game or Social Media Game?
A predictive engine for the NBA All-Star vote
Model inputs & outputs: x -> [Network importance scores, Standard & Advanced statistics], y -> [All-Star fan vote, All-Star player vote]
- allstar-votes:
- contains a series of html files pulled programmatically from basketball-reference.com
- parse.py, parse.sh, players-parse.py, usernames-and-handles.py: scripts to process all the html and associate entries for each player with a username
- allstar-votes/player-votes: collection of player names and number of player votes received for a given season. Some logic for processing csvs
- allstar-votes/votes: collection of player names and the number of fan votes receieved for a given season
- ig:
- series of csvs pulled from Instagram containing information like who a user follows, who follows them etc.
- ig/processed: the above, but processed (removed all redundant information)
- ig/scripts: the core logic for this project, including PageRank.ipynb which processes all follower information and generates importance metrics
- ig/scripts/count: logic for mimicking a browser and obtaining a follower count for each player (also an input to the model)
- kaggle:
- logic for making and running the learning model
- contains raw statistics (basic and advanced)
- kaggle/gs: logic for scraping and obtaining the missing "GS" (games started) statistic from bbref.com
- env: virtual environment
social-media-or-ball-game's People
Contributors
Watchers
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent