Articles, Blog

SciSports | Using AI to Calculate the Talent and Value of Soccer Players |SAS Customers

September 10, 2019

is a sports data analytics company. We help clubs with
scouting, and we’re also developing a real-time
tracking system, which means that we are placing 14
cameras alongside the pitch and we try to track
every move of a player with the goal of
improving the game. One of the added
values of BallJames, the system that
we’re developing, is that we want to also
retrieve the reconstruction. Which basically means
we want to generate more information from a
player, not only his X and Y positions on the pitch. Which you can’t do
without the 3D data is, OK is he scoring with his
left foot or his right foot, or is he scoring with
a header, how high was the jump that he just did? By bringing that 3D
aspect into the mix you get a completely
different class of new analytics that
wasn’t available before. What we do is we collect the
video, we encode the video, we stream it over
the network, we bring it to a central
unit within the stadium, send it to the cloud, and
then process all the analytics there. One of the things
that we want to do is use deep learning
approaches in order to do player detection, ball
detection, and the tracking part. What SAS can provide
us with ESP is that they can do very quick
inferencing of, let’s say, a trained deep learning
model we currently have. And by combining
this into Viya you have this in-memory platform,
so you can basically train your model in the
cloud or whatever, wherever your resources are. You can deploy the model
in-memory onto the camera, and then do the
inferencing in real-time. So it’s pretty exciting, and
if we can make this work, we can actually
bring this product to the market fairly quick. I mean time to market is the
most important factor in sports and sports data analytics. If we can bring
this to the market, it’s a completely
new tracking system that follows a completely
different approach than most of the competitors
that are out there. And also being able to
provide this new class of data and analytics. Yeah that’s going to
change the complete game.

1 Comment

  • Reply bet Analytic July 2, 2018 at 4:40 pm

    Hi, nice video. I'd like analyse a football match using SAS in order to predict the scores

  • Leave a Reply