SportSense: An AI-Powered Platform for Predicting Athlete Performance and Injury Risk

Abstract

Modern sport athlete management practices have evolved over time to be more quantitative, focused on improving athlete performance while considering their health. However, existing methods for analyzing athletes involve biased subjective criteria, lack of anticipation, and incomprehensible data analysis. The present paper presents SportSense, a simplistic solution that predicts injuries and assesses the preparedness of players. SportSense leverages machine learning algorithms like the Random Forest classifier, which evaluates data on workload, physiological conditions, and fatigue from pre-existing datasets. Training and prediction occur offline, followed by embedding the predictions in the web user interface. More specifically, SportSense comprises a web dashboard using the visualization framework Chart.js for visualizing information on risk profile, trends, and workload assessment. Additionally, a personalizable XAI solution using the manually designed TreeSHAP algorithm is presented to determine the crucial features used for making the prediction. In summary, SportSense improves athlete evaluations by detecting critical workload due to fatigue.

Introduction

The text discusses the growing importance of data-driven injury prevention in professional sports, where maintaining athlete health and performance is critical. Although large amounts of sports-related data—such as workload, match statistics, and performance metrics—are available, many teams struggle to use this data effectively. Existing machine learning solutions can predict injuries but often lack explainability and user-friendly visualization, making them difficult for coaches and sports staff to understand and apply.

To address these challenges, the proposed system, called SportSense, introduces a lightweight and explainable machine learning framework for predicting non-contact sports injuries. The system processes player-related data to generate features such as the Workload Index and Fatigue Index, which measure training intensity, distance traveled, game frequency, and rest periods. These features are analyzed using a Random Forest classifier to predict the probability of future injuries.

A key focus of SportSense is Explainable Artificial Intelligence (XAI). Using methods like TreeSHAP, the platform explains the factors behind each prediction, such as fatigue or insufficient recovery time, through visual graphs and textual insights. This makes the predictions easier for coaches and decision-makers to interpret and act upon. Machine learning operations, including feature extraction, training, and prediction, are performed offline using a Python pipeline, and the outputs are stored in CSV and JSON formats.

The platform uses a static web-based dashboard built with HTML, CSS, JavaScript, and Chart.js to visualize athlete profiles, injury risks, workload patterns, and performance metrics. By avoiding a server-side backend, the system remains lightweight, portable, and easy to deploy. All interactions, such as sorting and filtering, are handled on the client side.

The literature review highlights previous research in machine learning for sports injury prediction, blockchain applications, Web3 systems, and AI-based governance mechanisms. Existing studies often face limitations such as lack of scalability, class imbalance issues, insufficient explainability, weak user convenience, and poor real-time monitoring. SportSense aims to overcome these issues by integrating machine learning, explainable AI, and decentralized technologies into a single, easy-to-understand sports monitoring system.

Conclusion

In conclusion, the design and study of the SportSense project have brought to light the enormous transformation that occurs when artificial intelligence is introduced into the sports sciences field. By overcoming the difficulties encountered by the use of conventional methods in evaluating the performance of the athletes, a new technique for predicting injuries and monitoring the athletes\' performance levels has been devised. In this particular instance, the major aim was to employ artificial intelligence to create a prediction model using the random forest method. However, it should be emphasized that the implementation of XAI has been one of the main tools in bridging the gap between the results of machine learning and their application in coaching. This is because XAI allows coaches to learn the reasons for making a prediction, such as extreme fatigue, high loads during matches, or insufficient recovery time. Therefore, it is apparent from the results that coaches are more likely to accept evidence-based decisions when they understand the reasons for making a particular prediction. At present, SportSense being a static and web-based predictive analysis tool is considered to provide a very stable platform on which the effectiveness of SportSense can be evaluated. It is guaranteed due to the fact that the designed architecture of the tool allows running SportSense offline and visualizing its results in the software. At the same time, such an approach implies inability to adapt and update it in the process of use. Some of the future areas for research include the implementation of semi-dynamic and real-time systems in this case. The application of the pipeline system, which automatically performs the embedding of the data without the need for manual embedding of the data, is one such example. Other research efforts could also include the application of even more data from other sources. Besides, there could also be an improvement in the efficacy of the models based on real-world data. There could also be improvements in explainability and prediction models.

References

[1] Chen, X., Li, Y., & Wang, H. (2023). Healthcare Data-Based Prediction Algorithm for Potential Knee Joint Injury of Football Players. International Journal of Distributed Sensor Networks [2] C. Barton et al., “Injury Prediction in Competitive Runners With Machine Learning,” Int. [3] J L. Oliver, F. Ayala, and M. B. A. de Ste Croix, “Using Machine Learning to Improve Injury Risk Prediction in Elite Youth Soccer Players,” J. Sports Sci., 2020. [4] C. Leckey and N. van Dyk, “Workload Monitoring and Injury Risk in Sports,” Br. J. Sports Med., 2019. [5] E. Cust et al., “Machine-Learning Models for Injury Prediction in Team Sports,” J. Sports Sci. Med., 2018, doi: 10.1186/s40798-019-0202-3 [6] A. Rossi et al., “Sequence-Based Machine Learning for Injury Prediction,” J. Sci. Med. Sport, 2019. [7] B. Hulin et al., “Training Load and Injury: Acute :Chronic Workload Ratio,” J. Strength Cond. Res., 2016. [8] P. Zahariadis et al., “An Overview of Machine Learning Applications in Sports Injury Prediction,” 2023. [9] Y. Lu, A. Sperry, D. Shapiro, S. Stephan, and T. Pavlovic, “A Deep Learning Approach to Injury Forecasting in NBA Basketball (METIC),” J. Sports Analytics, 2021. DOI: 10.3233/JSA-200529. [10] S. Jauhiainen et al., “New Machine Learning Approach for Detection of Injury Risk Factors in Young Team Sport Athletes,” 2021. [11] Y. Lu et al., “Machine Learning for Predicting Lower Extremity Muscle Strain in NBA Athletes,” Orthop. J. Sports Med., 2022. [12] P. Reche-Soto et al., “Enhancing Sports Injury Risk Assessment in Soccer Using Machine Learning,” 2024.. [13] M. Baladaniya and A. K. Choudhary, “Artificial Intelligence in Sports Science: A Systematic Review and Meta-analysis,” J. Clin. Med. Kazakhstan, vol. 22, no. 3, pp. 64–72, 2025. [14] C. Vallio et al., “Artificial Intelligence in Sports Biomechanics: A Scoping Review,” Bioengineering, vol. 12, 2025. DOI: 10.1016/j.jbiomech.2020.110106. [15] P. Plisky et al., “Association of Y Balance Test Reach Asymmetry and Injury Risk,” Clin. Sports Med., 2015. [16] M. Luu et al., “Machine Learning Outperforms Logistic Regression to Predict Next-Season NHL Player Injury,” Orthop. J. Sports Med., 2020. DOI:10.1177/2325967120953404 [17] K. He and S. Zhang, “Prediction Model of Juvenile Football Players’ Sports Injury Based on Text Mining,” Mobile Information Systems, 2021. [18] A. Rossi, T. Perri, A. Cicchetti, and M. Marcone, “Predicting Athletic Performance from Physiological Parameters Using Machine Learning: Example of Bocce Ball,” 2019. [19] J. G. Cladino et al., “Current Approaches to the Use of Artificial Intelligence for Injury Risk Assessment and Performance Prediction in Team Sports,” Sports Med.–Open, 2022. DOI : 10.1186/s40798-019-0202-3 [20] A. Majumdar et al., “Machine Learning for Understanding and Predicting Injuries in Football,” 2022. [21] M. Choi, K. Lee, and K. Lee, “Injury Prediction in Korean Adult Field Hockey Players Using Machine Learning and SHAP-Based Feature Importance Analysis,” Appl. Sci., vol. 15, p. 8946, Aug. 2025, doi: 10.3390/app15168946

Copyright

Copyright © 2026 Mohammad Afnaan Anwar, Fouzan Taha Hussain, Syed Abdul Jaleel, Dr. K. Kalyani. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.