Parallel Session: Data driven sports research

Start 13.30, Wednesday 7/9


SB3-L111 (TRACKS environment). I suggest all participants gather with the chair (Moa Johansson) in the lunch area at 13.15 and then we will all walk over together so no-one gets lost at campus. After the talks there will be a guided tour of the new physiology lab where we have opportunities to use a wide treadmill, high-speed cameras and other sensors for various projects associated with sports.


13.30 - 13.45: "Taking the next step forward in running-injury research", Jonatan Jungemalm

13.45 - 14.00: "Machine Learning from Pacing Patterns for Half-Marathon", Johan Lamm + Johan Atterfors

14.00 - 14.15: "Automated route choices in orienteering", Marco Della Vedova

14.15 - 14.30: "Skisens AB - Unique field data and its interpretation", Johan Högstand

14.30 - 14.45: "Many source models for recovery and performance optimisation in sports", Johan Rogestedt

Around 15.00: Guided 15 minute tour of TRACKS physiology lab for those interested.

Abstracts of the talks:

Jonatan Jungemalm, Gothenburg University
Title: Taking the next step forward in running-injury research
Abstract: Running-related injuries (RRIs) are still a major barrier to maintaining a physically active lifestyle among recreational adult populations around the world. Despite intensive research, the proportion of injuries among runners has not been reduced over the last 30-40 years. Some explanatory factors for that include a lack of studies applying causal injury models as well as too small sample sizes – especially with regard to specific RRI diagnoses. This talk will briefly review some of the previous work on how to prevent injuries in running. In addition, an outline of how data-driven research can aid future RRI research will be made.

Johan Lamm + Johan Atterfors, Chalmers
Title: Machine Learning from Pacing Patterns for Half-Marathon
Abstract: Every year around 40-50 000 runners participate in Gothenburg Half Marathon, one of the worlds largest half-marathons. Our goal is to investigate if we can use historical public results data to analyse what is indicative of both good and bad performance, using techniques from data science and machine learning. We base our work on complete result data (over 400 000 entries) for ten years (2010 – 2019) where finishing times and 5km split times are recorded. Here, we present two machine learning models: One to predict finishing times of runners, and one for identifying which runners will hit the wall.

Marco Della Vedova, Chalmers 
Title: Automated route choices in orienteering
Abstract: The problem of route choice in orienteering is a perfect example of path planning in open terrain. Given an orienteering map and a course, which is the fastest route that can be deduced only from the map?
How does this route relate to the routes chosen by top orienteers in the terrain, given the GPS data from orienteering races?

Johan Högstand, Skisens AB
Title: Skisens AB - Unique field data and its interpretation
Abstract: Skisens AB, a sports tech company developing the world’s first powermeter for cross-country skiing, based on a unique sensor design and software platform generating high resolution activity data from multiple integrated sensors. The data collected from Skisens is stored in different levels of abstraction, as a challenge with new data it is high focus on detecting the key parameters and its interpretation for the user in a useful way as well as unlocking the potential with big data analysis. Johan will introduce Skisens data streams and how new and more data enables interesting research projects.

Johan Rogestedt, Svexa
Titel: Many source models for recovery and performance optimisation in sports
Abstract: Today, the availability for devices measuring performance and recovery related metrics is rapidly increasing. By harmonising data and working with models using many sources in the intersection between domain knowledge and modern modelling tools, we have a better chance to create truly useful insights to coaches and athletes.

Page manager Published: Mon 05 Sep 2022.