Data Preparation for Injury Prediction
Posted by Mladen Jovanovic on 23/08/2018 in Monitoring, Non-Membership Content, Statistics & Data Analysis
4 Comments

Data Preparation for Injury Prediction

I have recently wrote a technical note (actually a video) for Sport Performance and Science Reports journal regarding Data Preparation for Injury Prediction. Both data and R core are available on GitHub repository. The purpose of this technical note is to showcase one particular way of preparing (and engineering) data for injury prediction tasks, as well as start the discussion with fellow sport scientist. As always, any feedback is highly welcome.

 

Tags:, , , , , ,
I am a physical preparation coach from Belgrade, Serbia, grew up in Pula, Croatia (which I consider my home town). I was involved in physical preparation of professional, amateur and recreational athletes of various ages in sports such as basketball, soccer, volleyball, martial arts and tennis. Read More »
free-memeber-button
free-memeber-button

Related Articles

Welcome to Complementary Training Community! › Forums › Data Preparation for Injury Prediction

This topic contains 4 replies, has 3 voices, and was last updated by  Mladen Jovanovic 4 weeks ago.

Viewing 4 posts - 1 through 4 (of 4 total)
  • Back to forum
    Posted in: Blog Posts and Articles
  • Marco Porta
    29/08/2018 at 14:18 #23629

    Wonderful article.
    I will try to implement this approach with me team.
    One thing: on your opinion, which will be the best solution to implement gps data if I have only 10 gps? I will use gps data only on some athlete (that are several every training) or it will be better to use average values and apply them to the whole team?

    Thank a lot again!

    Marco

    Mladen Jovanovic
    04/09/2018 at 09:43 #23660

    Hi Marco,

    Glad you found it useful. Regarding only having 10 units – I think the BEST use might be to create a list of exercises/drills you do and have a certain ranges of values (e.g. HSD per min od drill, # Hard Acceleration per minute of drill, etc). You can rotate athletes to get a glimpse, as well note the position of the athletes (to also see if there are any differences).

    Having this drill database makes you more ‘precise’ in ‘predicting’ the load. This is by far not perfect, but beteer than nothing.

    As for injury prediction with only 10 units? Hmm… Having 30 unitsts, injury prediction is still pipe dream, so don’t need to worry about it. What can be done is to use the drills database to ‘estimate’ individual loads (using position and minutes played in a given drill) or as you say ‘average’ (although then you are predicting for a TEAM, not individual).

    Hope this all make some sense 🙂

    Glenn Madison
    19/08/2020 at 00:28 #29355

    Hi Mladen,

    Would you be able to help me design a regression model for GPS data? I work in non league so we don’t have GPS Units and I want to predict HSR etc from SSG based on data collected in the literature. I want to be able to manipulate the pitch sizes and durations to then get a prediction of the external load on the players.

    Any help is appreciated!!!

    Mladen Jovanovic
    22/08/2020 at 09:43 #29393

    Hi Glenn,

    Might be hard to do that. But let me make sure that understood this correctly – you do not have GPS, but want to ‘predict’ HSR from SSG data? Ok, what data do you have? From literature, you can only get ‘expected’ values (i.e., mean).

Viewing 4 posts - 1 through 4 (of 4 total)

You must be logged in to reply to this topic.