Data Science

TLDR;

This module introduces data science, highlighting its multidisciplinary nature and applications within human movement science. It explains how data science combines computer science, mathematics, statistics, and domain-specific knowledge to extract valuable insights from complex datasets. The module also covers the four pillars of data science: domain knowledge, mathematics and statistics, computer science, and communication and visualisation. It further classifies biomechanical data into qualitative (nominal and ordinal) and quantitative (continuous and discrete) types, and explores the applications of data science in gait analysis, sensor data processing, fall prevention, health monitoring, neurological disorder analysis, and biomechanical analysis.

• Data science is a multidisciplinary field combining computer science, mathematics, statistics, and domain-specific knowledge.
• The four pillars of data science are domain knowledge, mathematics and statistics, computer science, and communication and visualisation.
• Biomechanical data can be classified into qualitative (nominal and ordinal) and quantitative (continuous and discrete) types.
• Data science has various applications in human movement science, including gait analysis, sensor data processing, fall prevention, health monitoring, neurological disorder analysis, and biomechanical analysis.

Introduction to Data Science [0:00]

Data science is a multidisciplinary field that integrates computer science, mathematics and statistics, and domain-specific knowledge. The intersection of mathematics and domain knowledge forms traditional research, while computer science and domain knowledge overlap in software development. Machine learning arises from the combination of computer science and mathematics/statistics. Data science encompasses all three, using this combined knowledge to extract insights from data. Various algorithms and tools are designed to analyse large, complex datasets, aiding informed decision-making and solving complex problems efficiently.

Four Pillars of Data Science [2:09]

Data science is underpinned by four main pillars: domain knowledge, mathematics and statistical skills, computer science knowledge, and communication and visualisation skills. Domain knowledge, particularly in human movement analysis, requires understanding biomechanical data types, collection methods, biomechanical models (e.g., how the skeletal system produces movement with muscular force), and biomechanical parameters like joint angles and forces. Mathematics and statistics necessitate knowledge of linear algebra, calculus, descriptive statistics (mean, mode, median), and inferential statistics (regression equations) for data analysis. Computer science involves understanding big data technologies for data storage and programming skills to implement mathematical computations. Finally, storytelling skills and visual art design are crucial for presenting data effectively to diverse audiences, connecting different domains, and conveying information clearly through visuals.

Application of Four Pillars in Sports Science [6:39]

In sports science, the four pillars of data science are essential for analysing and improving athletic performance. Domain knowledge involves understanding movements and parameters like joint angles, positions, velocities, and accelerations. Mathematical skills, such as linear algebra and calculus, are used to calculate velocities and accelerations from biomechanical data. Computer science is crucial for managing and extracting data from athlete management systems. Effective communication and visualisation skills are needed to present findings to athletes, coaches, and sports scientists, ensuring the information is understood and can be applied to enhance training and performance.

Types of Biomechanical Data [8:11]

Biomechanical data is broadly classified into qualitative and quantitative data. Qualitative data, which cannot be measured numerically, includes images and text, and is further divided into nominal and ordinal data. Nominal data, also known as categorical data, includes variables without any order (e.g., gender, nationality). Ordinal data has a sequential order but does not allow for arithmetic functions (e.g., feedback scales from 0 to 10, education level). Quantitative data, which can be measured numerically, is divided into continuous and discrete data. Continuous data can be expressed in fractions and decimals (e.g., kinematic data like joint angles), while discrete data consists of distinct, separate values represented in whole numbers (e.g., the number of players in a competition).

Applications of Data Science in Human Movement Science [12:20]

Data science significantly enhances various applications within human movement science. In gait analysis, data science tools and techniques can recognise and classify different gait patterns, aiding in diagnostics and treatment planning, particularly in orthopaedics. For data collection and sensor technologies like motion capture systems and wearable devices, data science automates collection procedures and synchronises data for meaningful insights. In fall prevention for the elderly, data science identifies indicative patterns of falls, enabling the development of preventive interventions. Health monitoring and wearables benefit from data science by processing and analysing movement patterns and sleep quality to provide valuable insights into overall health. Neurological movement disorder analysis uses data science to identify biomarkers for early disease detection and to monitor rehabilitation processes. Biomechanical analysis applies machine learning and statistical methods to uncover patterns and correlations in biomechanical data, improving efficiency and reducing errors in understanding complex movement variables.