Contract negotiations in the NHL can be a lengthy, complex process, with General Managers trying to sign players for as little as possible while a player's agent aims to sign their client to the biggest contact they can. But which side's determination of a player's worth is the right one? Using machine learning I aim to resolve situations such as this by building a model that can output a salary based on 50 common and advanced player statistics. The ability to capture the value of a players production in dollars adds more transparency, efficiency, and consistency to how contracts are determined. As a result, both sides benefit: players receive fair pay proportional to their output, and General Managers get the best value for their pay. Fewer bad contracts thus increases a team's ability to assemble a better on-ice product, which benefits the players, Management, and the fan base.

1. Data

• Abbreviations and Definitions

• Raw Data

• Cleaned Data

  -Contract data obtained from www.capfriendly.com

  -Statistical data for players between 2007 and 2020 obtained from www.hockeyreference.com

2. Documents

• Final Report
• Presentation

3. Notebooks

• Data Wrangling
  • Skaters
  • Goalies
  • Contracts
• Exploratory Data Analysis
• Preprocessing and Modeling