Science Library | ODY Library | ISEI | SLU Home

 

 

 

 

Travis Gingras '06 Wins the 2006 Robert R. Churchill Memorial Prize
for Student Scholarship in GIS

Travis Gingras, '06, a math major and member of the club hockey team, has won the NITLE 2006 Robert R. Churchill Memorial Prize for Student Scholarship in GIS. Travis' submission and senior thesis is Application of a Geographical Information System to Ice Hockey Shots.

As many hockey enthusiasts know, it's not necessarily what type of shot or how fast the shot was going but rather where on the ice the shot was taken from, this is because you can find angles where goalies are weaker. The obvious beginning step for this project was to create a map where the shots could be placed. Travis georeferenced both the NHL standard hockey rink and goal net to their respective dimensions.

 

By watching game DVD's Travis located each shot taken and plotted a point on the georeferenced rink and the appropriate spot on net. Shot details, including outcome of the shot, shot type and player ID were then entered into the geodatabase. Three types of database variables were created: 1) Coded Variables that change from shot to shot (ex: type of shot); 2) Coded Variables that change from game to game (ex: opposing team); and 3) Long Integer variables, which have only limited possibilities and can be entered later (ex: opposing goalie's number).

Coded Value	Description
1	Slap Shot
2	Snap Shot
3	Wrist Shot
4	Backhand
5	Deflection
6	Miscellaneous

Coded variables allowed for the rapid entry of data into the geodatabase. The system we have created allows a single person to track multiple events of the game, albeit not during the game, with tremendous accuracy and very little time commitment. Data can be displayed as a single game, multiple games, by an individual player or shot type.

This powerplay map shows us a very bold statement; 40 of the 73 attempted shots were taken on the powerplay. There were only 8 penalties for Brown, which means that all 40 shots came in less then 16 minutes of the 60 minute game. By pooling data from several games Travis was able to analyze shooting and scoring patterns during and between games using spatial statistics such as point density and Kriging.

 

The point density map was calculated for all the shots in the "Offensive Totals" game file. Where the map output is darker blue more shots are likely to come from there. This can be useful when tracking the tendencies of a team or how an individual player shoots over many games. The goal is to create a continuous map/surface to indicate where it is more likely that the shot will be blocked, or where it is more likely that a goal will be scored.

 

Hockey represents a confined area with the possibility of quick changes in values. Kriging takes into account the global trend of the data, in this case the mean of all the data placed on the rink. This method generates a more accurate surface than other methods such as Inverse Distance Weighting (IDW). Below we can see a representation of what a Kriging map would look like for Slap Shots.

These results have been compiled in a paper co-authored by Travis, Dr. Robin Lock, Burry Professor of Mathematics, Computer Science & Statistics, and Chris Wells, Associate Men's Hockey Coach. The paper will be presented this summer in a "Sports and Statistics" session at a national statistics conference. Travis acquired his GIS skills through the St. Lawrence GIS Program where he has been a GIS technician for the past three years. Information about the Churchill Prize and NITLE's GIS Initiative is available at the NITLE website.