Haven't you ever been in a debate about who was a better, more clutch player? Was it Michael Jordan, or Lebron James? Was it Steve Young, or Joe Montana? Who threw more 4th quarter touchdowns in big games?
No more arguing. Let's look at what the data tells us.
How it works
At the biggest moments in the game, some players come up big, and others fall short. Those that come up big will have a strong clutch rating, and those that fall short will have their overall clutch rating reduced. The dataset we've used contains play and game-time data for NFL players from 2009-present.
An overall clutch rating is computed through the aggregate of all plays by that player, over the entire dataset. If you want to see what plays went into the equation, you can click on a player and get a more detailed view -- which will include things like height and weight data, in addition to a list of all of the plays, along with each individual play's clutch rating (can be positive or negative), which is factored into the player's total score.
The real kicker is that we give users a video link to watch a particular play they are interested in.
Challenges we ran into
Clean data sets with strong historical data were hard to come by.
Accomplishments that I'm proud of
With no UI folks on our team, we think we did a pretty good job. The tables also have search functionality, and the additional feature of having a link to play-videos is exciting.
What we learned
Look out for semi-colons, they'll getcha every time.
What's next for Clutch Ratings
We plan on launching the site.