Ideas.

[chriskuz]

Place to write down thoughts on how the project could scale.

[chriskuz]

• Opposing team goalie analytics that looks at how the goalie has made saves this year. Incorporates analysis of mobility patterns of Rangers on offense, goalie vulnerabilities for most errors made from different parts of the ice, and other patterns.

• Goaltender analysis. Observing by the numbers of different goaltenders this season.

• Offensive forward analysis for top players and comparison across the league. Looking at particularly Artemi Panarin and Mika Zibanejad comparing across the league.

• Observation and possible associative analysis on the age of the team and penalty minutes discovered. Could even try implementing an ML model to predict penalty minutes with various attributes across players. Supervised learning comparison of previous seasons to this season.

[chriskuz]

• global_match_history_report.ipynb serves as the report file for analyzing teams' historical games against one another. This is updated nightly on moneypuck and is a good place to compare seasonal statistics. ipynb might be best to use, but a python OOP script might be best for repetitive/nightly updates...makes our lives easier with utilizing custom tools.

[chriskuz]

• one alternate idea is to not give that much power to the user where all entry parameters are defined by them. It may be better to have functions and methods for players and static team statistics to be outputted. Then again, every player/team is different, so it may be easier to still have the option to set your own parameters. Perhaps if the module will eventually be broken up into offensive analysis and defensive player analysis and a spare tool of user-input analysis. This might be the best option to create a very deep module.

[chriskuz]

• analyze Rangers power play effectiveness. Look into all important details.

[chriskuz]

• allow a customizable dashboard property where multiple plots can be prepped together at once. Preliminary dashboard designs can be player specific in which plots are prepped side by side to one another. These can be template based. However a dynamic dashboard with adding plots side by side would be favorable.

[chriskuz]

Project Structural Idea:

Main script --> initializes functions --> asks use for inputs on what to do

Answer on what the user wants will dictate what kind of search will be performed (whether web scraper or no web scraper is needed) --> potential further detailed input credentials --> begin scrapes/source code

Post program run:

• use has option to clear cache
• use can analyze another plot