Finding best team composition on a particular ground and making a team based on statistics obtained from the dataset.
Install Selenium in your working environment using- pip install Selenium Download the Chrome driver attached Install pyfpgrowth package using- pip install pyfpgrowth
- Using the datasetyears.py file, team compositions of winning teams at Wankhede Stadium were imported into a csv file:teamcomposition.csv. The data was scrapped from the Cricbuzz.
- Using the fp growth algorithm, the combination of -batsmen, bowlers, batting-allrounders and bowling-allrounders was found
- Web scraping was also used to find the current squad of MUmbai Indians from howstat.com
- Ball by ball data available in deliveries.csv was used.
- To find player statistics only last 2019, 2018, 2017 seasons were considered.
- Players were clustered into categories-Batsman, Bowler, Bating-allrounder, Bowling-allrounder
- By keeping in mind the winning combination, team was constructed.
- Players that have not played a single match were not considered as data is not available. Getting data of domestic matches and using a factor to adjust the difficulty can be done.
- The clusters formed could be improved.
- Different approaches to filling the 11 can be employed.