A minimalistic toolbox for extracting features from sport activity files
Iztok Fister, Luka Lukač, Alen Rajšp, Iztok Fister, Luka Pečnik, Iztok Fister
Abstract
Nowadays, professional, as well as, amateur athletes are monitoring their sport activities/training using modern sport trackers. These devices allow athletes to capture many indicators of sport training, e.g. location of training, duration of training, distance of training, consumption of calories. Until recently, not enough devotion was given to those indicators that are not visible directly, but can be obtained as the result of extensive data analysis, e.g. information extracted from topographic maps, weather conditions, and interval data. In line with this, the present paper is dedicated to describing the new toolbox for extracting features hidden in sports activity files. The results of the extraction serve as entry points for deep data analysis, that allows us to build intelligent systems for training support.