I've always been interested in the developing world of hockey analytics and this was my first project in the field. I grouped NHL teams based on the average amount of goals they scored and allowed in a season. Click the github button to access the full program.
I built a C++ program that went over all teams from all seasons for the past decade and grouped them into one of the four above categories. The program then went and analyzed every game played from each year, determined which category each team was in in each head to head matchup, and recorded how each category faired against the others, which I measured in winning percentage.
The figure below represents how the data was stored. Each year was linked through a binary search tree, and each year pointed to a hash table, which had four containers to store all four categories for the teams.
Unfortunately, the results weren't statistically significant enough to conclude that grouping teams in this way gave any insight to who would win in a given matchup. The results can be seen below.
