2 Evaluating players

This is a graded exercise. Deadline is strict.

It is summer of 2018. The head scout of your club (one of the biggest in the Premier League) wants to scout for a particular position from anywhere in the top 5 leagues using Wyscout data.

Group 1: Defender, Manchester United

Group 2: Forward, Arsenal

Group 3: Midfielder, Tottenham Hotspur

Group 4: Goalkeeper, Manchester City

Group 5: Defender, Liverpool

Group 6: Forward, Chelsea

He is interested in hearing about a variety of methods for scouting players, so he asks you to try out one of the methods from the course or even one of your own. You should primarily use the Wyscout data, but can also use Transfermarkt to find the prices of players in summer 2018.

You should do the following steps:

1, Implement one of the methods (plus/minus, percentiles, Markov chain, possession chains or one of your own) on the data. Decide on a suitable metric for ranking players and make a top-10 list of players in your position for the whole league.  Write a simple non-technical text (half a page) explaining your metric to the scout and what assumptions it makes. Provide a separate runnable piece of code (preferably in Python) which when put in the directory of the Wyscout data creates the top 10. This code must be single file and only uses standard libraries (mplsoccer is fine) (10 points)

2, Use your metric to find a single player in another league (not the Premier league), who you would recommend signing. Add additional statistics and visualisations to explain the strengths and weaknesses of that player (these can use World Cup data where appropriate). Create a two page report, in a poster style, with as many visualisations as you want but max 2 pages on that player .  (5 points)

3, On Thursday the 30th September, you will be asked to share and present your player for 2 minutes within your group. After the presentations, you will have a group discussion comparing your choices. You should write a single page contrasting your own and the other reports and make a final recommendation on this basis (participation in meeting 2 points; final report 3 points)

The final report should be 4 pages: 1 page for part 1, 2 pages for part 2, and 1 page for part 3. It should include a separate, runnable file. Important: this code should be a single file run immediately if placed with in the same directory as the Wyscout folder, which in turn contains Wyscout/events etc. i.e. exactly as is done in the expected goals tutorial.