6 136 CHAPTER 6 GENERAL DISCUSSION This thesis aimed to reduce the data analytics gap represented by the four identified problems while examining the associated potential solutions to enable meaningful insights and predictions related to physical activity and physical performance. These insights and predictions can be used to make more informed decisions regarding physical activity and physical performance interventions. The chapters in this thesis explored potential solutions to the identified problems. We briefly discuss the problems and their corresponding potential solutions to revisit the introduction. One problem in data analytics for physical activity and physical performance is the limited use of personalised data for meaningful insights and predictions. The potential solution is using personalised data from wearable sensor devices like Fitbit or Garmin or optical tracking systems such as SportsVU. Another problem is the vast amount and complexity of data, making it challenging to build accurate prediction models. The suggested solution is to use more sensitive performance measures in combination with various machine learning algorithms. The third problem involves overly simplified models and assumptions that make unrealistic assumptions about the underlying reality, limiting the value of insights. The possible solution is a causal roadmap combined with a causal model. The fourth problem is the absence of confounding variables, such as contextual variables or individual characteristics, that could influence physical performance. The two suggested methodological solutions use statistical methods that account for the absence of confounding variables and a two-way approach that applies traditional statistical analysis and machine learning to the same incomplete dataset to identify the best fit. All the suggested solutions could reduce the data analytics gap in physical activity and physical performance data and provide meaningful insights and predictions. KEY FINDINGS AND DISCUSSION The use of personalised data enables personal meaningful insights and predictions As found in Chapter 2, based on personal Fitbit step data, we could predict whether a person would reach her/his daily number of steps. Similarly, in Chapter 3, using personalised SportsVU data enabled predicting individual soccer players’ performance throughout a soccer match. However, it is unclear upfront which level of granularity in the data is needed to enable meaningful insights and accurate predictions. In some cases, a higher level of granularity
RkJQdWJsaXNoZXIy MjY0ODMw