In the realm of sports science, accurate and reliable workload data is paramount to understanding and improving player performance potential.
However, there are instances where workload data may be missing or incomplete, posing challenges for those professionals charged with the longitudinal monitoring of athletes.
In this article, Zone7’s Performance Director, Rich Buchanan explores how the company tackles clients’ missing workload data and the methods that Zone7 employs to ensure longitudinal data integrity.
Missing Match Data
In the relentless world of football and similarly in other sports, it is not uncommon for players not to wear tracking devices such as GPS units on match day. This can happen for various reasons and creates a challenge for performance staff.
The second best source for match data would be from optical tracking systems. If these are not available, Zone7 will apply a proprietary data simulation algorithm to ‘fill in the gaps’ based on other physical match data parameters. The generated parameters will be consistent in ranges and types to the specific team’s GPS units, allowing for continuous longitudinal monitoring and workflow enablement.
Missing Training Data without Reliable Alternative Source
In situations where players do not wear GPS devices during training or there is no reliable alternative data source like optical tracking available, Zone7 collaborates with the team to obtain referenceable attributes regarding the training session. This information serves as the basis for simulating an appropriate training session in accordance to the team’s training microcycle based on the player’s historical data. While not necessarily a perfect estimation, this approach has been proven valuable (relative to the alternative of no data) to creating continuity in the human analysis process and the digital modelling process.
Missing data from International Duty and Missing Workload GPS Data
When players are away from the club environment on international duty, the preferred path is to integrate GPS provided by the international team. When this data is available, Zone7 will ‘translate’ this data into the club’s set of GPS parameters using a complex and tested mathematical transposition algorithm.
However, when international duty and workload data is not made available to the club by the international team, Zone7 employs a two-step approach to simulate players workload data from matches and training sessions for seamless longitudinal monitoring.
Firstly, match workload is estimated based on available data about the national team matches, using the player’s historical data as a reference, and by applying the methodology described earlier for simulating missing match data. Secondly, when possible, novel methods are used to estimate training data, often in collaboration with the national team. This approach aims to ensure that the data retains its completeness as much as possible, while quantifying and minimising the artificial “noise” that can be associated with various data simulation methods. If subsequent GPS data is made available at a later date by the international team, this can be backfilled to replace the simulated data.
Overriding GPS Data with Simulated Data
To address the issue of GPS device malfunctions and anomalies in the GPS dataset, which can lead to an individual player having inaccurate data, Zone7 applies various best-practice anomaly and outlier detection methods to identify and rectify erroneous data points. Once the data is cleansed, the missing values are simulated on the methodologies described in previous sections of this article. This combination of outlier detection and data simulation helps maintain data consistency and accuracy.
To ensure match data accuracy, when both wearable and optical datasets are available, a comparison is made between the data sources. If there is a significant discrepancy, Zone7 investigates the data to determine which data source is most usable and reflective of reality.This meticulous validation process guarantees that the final data accurately represents the player’s workload output on the pitch.
Data integrity tracking – the missing link
To know when to deploy these data solutions and track their accuracy, Zone7 includes a robust and scalable data integrity platform. This means that every bit of data coming in (e.g. from GPS or Optical tracking) is monitored for consistency so data gaps are detected in real time. In addition, the values themselves are continuously tracked so that any anomalies related to device issues of threshold or zone configurations can be detected and dealt with according to the client-specific policies.
Diving deeper
Many of the algorithms mentioned above are proprietary creations of the Zone7 team and are often customised per client scenarios. If you are interested to learn more about how these algorithms operate, the specific inputs and outputs and the rigorous validation process applied, please contact our team directly at info@zone7.ai.
Conclusion
Missing data can pose challenges in the field of longitudinal athlete monitoring, but Zone7 has devised innovative methods to overcome these obstacles. Employing a combination of data simulation algorithms, micro-cycle analysis, sports science expertise and collaboration with teams to handle missing data in football. This approach ensures data completeness, accuracy, and alignment for longitudinal athlete monitoring.
By leveraging the available data sources and utilising advanced artificial intelligence algorithms, Zone7 empowers performance professionals from a data perspective to inform on player management strategy and decisions.
Ready to Elevate your Performance? We provide a comprehensive suite of products and services utilizing data and AI to enable greater performance and durability amongst high-performance teams.
To find out more, Book a Demo today.
