& 149 SUMMARY used in behavioural, human movement, or sports science. Additionally, using Artificial Intelligence and machine learning based on wearable and monitoring systems data in sports is still in the early stages. While data analytics offers opportunities, problems in extracting meaningful, personalised insights and realising predictions based on physical activity and physical performance data exist. Four problems are identified in data analytics of physical activity and physical performance data, creating a data analytics gap. To address these problems, we propose potential solutions for each identified problem. The first problem is the limited use of individualised prediction based on personalised. In order to provide individualised insights and predictions, a boundary condition is that the data contains sufficient personal information. The potential solution is to use personalised data from wearables like Fitbit or Garmin or optical tracking systems such as SportsVU for individualised insights and predictions. The second problem is the vast amount and complexity of data, making it challenging to create accurate prediction models. The suggested solution is to use more sensitive physical performance measures in combination with various machine learning algorithms. These more sensitive measures can better detect changes in the measured system. The third problem involves the use of simplified models of reality and hypotheses that make unrealistic assumptions about reality, limiting the value of insights. The possible solution is to apply a causal roadmap combined with a causal model. The causal roadmap strictly prescribes which steps must be taken in analysis and clarifies the underlying assumptions. A causal model provides insight into how reality is modelled, and which variables have been left out. The fourth problem is the absence of variables in the data that can influence the measured physical activities and performance. For example, in soccer, contextual variables such as a home or away game or the weather. There are two potential methodological solutions. The first solution uses statistical methods that take into account the absence of influencing variables, and the second tests whether traditional statistical analysis or machine learning applied to the same incomplete dataset provides better insights and predictions. The aim of this thesis is to reduce the gap in data analytics by addressing the four identified problems. By exploring the associated potential solutions, it may become possible to provide meaningful insights and predictions regarding physical activity and performance. These insights and predictions can be used to make more informed decisions regarding physical activity and physical performance interventions. The chapters in the thesis explored the potential solutions to the identified problems. In Chapter 2, we used personalised data captured by wearable devices to predict employees’ daily physical performance automatically. Specifically, we used Fitbits to track the Hanze University of Applied Science employees’ daily step counts and employed
RkJQdWJsaXNoZXIy MjY0ODMw