Uma arquitetura e implementação do módulo de pré-processamento para biblioteca PyMove

Abstract

Trajectory data volume has increased greatly in recent years. Advances in location data collection technologies, an increase in the number of devices that use these technologies, and a reduction in the cost of these devices have all contributed to increased data volume. Trajectory data helps to develop work from different areas of knowledge. However, in the literature there are not many tools whose main focus is the preprocessing of trajectory data. The present work has as main goal of designing an extensible and flexible library to assist in the analysis of trajectory data, focusing on data preprocessing operations. The library is called PyMove and it has been developed in Python language, it supports the following stages of data preprocessing: noise filtering, stay point detection, compression, map-matching and segmentation. PyMove has the following advantages over other related work: It is a flexible library and supports more data preprocessing activities. However, PyMove has few operations for data compression.