Exploring new datasets can be challenging. Addressing this challenge, there is a whole field called exploratory data analysis that focuses on exploring datasets, often with visual methods.
Concerning movement data in particular, there’s a comprehensive book on the visual analysis of movement by Andrienko et al. (2013) and a host of papers, such as the recent state of the art summary by Andrienko et al. (2017).
However, while the literature does provide concepts, methods, and example applications, these have not yet translated into readily available tools for analysts to use in their daily work. To fill this gap, I’m working on a template for movement data exploration implemented in Python using MovingPandas. The proposed workflow consists of five main steps:
- Establishing an overview by visualizing raw input data records
- Putting records in context by exploring information from consecutive movement data records (such as: time between records, speed, and direction)
- Extracting trajectories & events by dividing the raw continuous tracks into individual trajectories and/or events
- Exploring patterns in trajectory and event data by looking at groups of the trajectories or events
- Analyzing outliers by looking at potential outliers and how they may challenge preconceived assumptions about the dataset characteristics
To ensure a reproducible workflow, I’m designing the template as a a Jupyter notebook. It combines spatial and non-spatial plots using the awesome hvPlot library:
This notebook is a work-in-progress and you can follow its development at http://exploration.movingpandas.org. Your feedback is most welcome!
- Andrienko G, Andrienko N, Bak P, Keim D, Wrobel S (2013) Visual analytics of movement. Springer Science & Business Media.
- Andrienko G, Andrienko N, Chen W, Maciejewski R, Zhao Y (2017) Visual Analytics of Mobility and Transportation: State of the Art and Further Research Directions. IEEE Transactions on Intelligent Transportation Systems 18(8):2232–2249, DOI 10.1109/TITS.2017.2683539