Free Data Scouting Tool
Matthias Werner
I turn disconnected processes into data-backed growth engines | RevOps | BizOps | Data & Analytics
How to build your own self-service football intelligence
Data scouting is more and more becoming an integral part of modern scouting departments in football clubs. Providers of various solutions to make data easily accessible, prepare it in an appealing way and enrich it with sophisticated models are increasingly establishing themselves on the market. Since I strongly believe in the power and potential of Data, Analytics and Business (Football) Intelligence, I was wondering if it is possible to create a simple data scouting tool based on publicly available data sets.
Is it possible to create a data scouting tool based on publicly available data at no cost?!
I would like to share my answer to this question with all interested parties. Here is as my free data scouting tool!
Instead of a long article about the importance of data and its analysis in professional football, I would just like to give a short introduction to my DIY tool.
A few remarks in advance:
The tool contains data of the big 5 leagues from the 2019/2020 season for a simple reason: For these leagues the most extensive public free data exists and further seasons would just have gone beyond the scope of this little experiment.
But in principle, the tool can be fed with additional data at any time. Be it with additional leagues or additional seasons of the Big 5.
Let's explain the handling of the tool with a concrete example:
You were asked to find a young midfielder for your ambitious club, who has already played in a top league for a few minutes and has solid defensive skills and potential in offensive play. Some may call him a box-to-box player.
First of all, we can translate the basic requirements into actual numbers and set our global filters accordingly:
- There is no focus on a certain league or player’s nationality, so we leave these filters untouched
- We select all “MF” positions to loosely narrow down to midfielders
- The player we are looking for should have played at least 450 minutes in the last season
- And the player should not have been older than 23 at the beginning of that given season
These are our global filters, our data set is already much better to oversee. Each dot represents one player.
Here should be a video, but unfortunately I noticed that it is not always displayed correctly. If it is not visible for you, you can follow this link.
?? Here is your direct way into the tool: Wow!
Now, let’s dive into more qualitative metrics. The scatter plot allows us to look at three different metrics at a glance. You can easily select which dimension you want to see on the axes and set a metric being reflected by the bubble sizes.
Let’s start with something basic. On the Y axis we want to see the number of tackles per 90 minutes, on the X axis the percentage of completed passes and the bubble size shows us the assists of the player in that given season.
That way, we can easily depict the most interesting players in the upper right quadrant. These players complete more passes than the average while being also active in defensive actions. Additionally, we can judge from the bubble size, that the bigger dots may create the most dangerous offensive actions.
Hovering over the dots reveals the underlying player and some additional base information.
Here should be a video, but unfortunately I noticed that it is not always displayed correctly. If it is not visible for you, you can follow this link.
While this is nice-to-have the magics start once you compare players in even more than just three different dimensions. To do so, let’s navigate to the 2x2 scouting matrix.
In the same way as in the first visualisation you can now achieve the fourfold gain in knowledge. Just fill in the axes of the charts with the metrics you are most interested in. If you then select a player in one of the four charts he will turn red, so you can easily see how he has performed in the other dimension.
For our test task, I select the following metrics:
Bubble Size: Minutes Played
Chart 1
Y Axis: Passes Attempted
X Axis: Passes Completed %
Chart 2
Y Axis: xG
X Axis: xA
Chart 3
Y Axis: Tackles
X Axis: Tackles Won %
Chart 4
Y Axis: Passes Completed into Final 3rd
X Axis: Key Passes
Here should be a video, but unfortunately I noticed that it is not always displayed correctly. If it is not visible for you, you can follow this link.
You can now just click some promising dots to discover the hidden gems in the data. Despite a couple of obvious marquee-players you may find some less apparent mates. In our example, Marc Roca could be such a guy, as he is above average in each of the metrics we compared. Seems like the tool is working, read the confirmation here ?? . Of course, you may argue, this player is not completely unknown, but this case is only meant to explain how the tool works, I am curious to learn which rare finds you get.
?? Here's the direct link to the 2x2 data scouting matrix: Click here
Now, I hope you will have some fun while data scouting yourself. Please let me know which interesting findings you discovered and in general if you like to tool and what should be improved.
Of course I am aware that this tool has a lot of shortcomings and of course it would not stand any comparison with a professional software. I just want to show that even with little (or no) money and a little effort and passion it is possible to create useful tools.
Looking forward to your feedback!
?? A few practical hints:
- The tool works best on desktop
- I just tested it with Google Chrome browser and 80% zoom, if you’re using a different browser optimal preferences may differ for you
- If you select a player a tooltip shows up with basic information and a link to the full player profile on worldfootball.net some of these links might be broken due to different spellings, I am super sorry for that!
?? Here is the link to the complete tool: Have fun!
In case you liked it, feel free to leave a comment or have a look at my other articles:
Startup FC - The opposite of what we have always done
How to make smarter decisions in football clubs
I turn disconnected processes into data-backed growth engines | RevOps | BizOps | Data & Analytics
4 年Seems like the video only work while being in private/inkognito mode ??
I turn disconnected processes into data-backed growth engines | RevOps | BizOps | Data & Analytics
4 年I just noticed, that for some reason, the video snippets are not always shown in the article. In case you missed them, I updated the article with links to the videos.