Advanced methods for processing and visualization of geospatial data
Academic Year: 2021/2022
Supervisor: Hruška Tomáš, prof. Ing., CSc.
Department: Department of Information Systems FIT BUT
Today's increasing requirements on data processing and visualization emphasize the importance of developing high-quality, quick, and user-friendly tools used for this purpose. Common types of visualization media (charts) are usually not sufficient to visualize complex multidimensional data. An example of such data is the geospatial data representing the relation of data with a geographical location (such as occurrences of some effects on the Earth's surface). The need to visualize such data goes beyond the academic areas of interest. We can meet with the geospatial visualization in the industry (distribution of bureaus, network topologies), public sector (public transport schemes), media (spread of disease, economic and demographic indicators), but also among ordinary users (private routes tracking). Those visualizations are usually arranged in advanced screens (e. g., dashboards).
Users can choose between two types of tools for geospatial visualization:
- Authoring systems: they allow users to create their custom thematic maps. Examples of such a tool are the systems used for general data visualization (e, g., Tableau or Grafana). Their functionality usually limits the usage of such systems. Hence, it is possible to use the solutions specialized in visualization only (e. g., Mapbox). Such systems usually provide advanced possibilities to create custom thematic maps without coding knowledge. However, learning such systems usually exceeds the knowledge of typical users. Moreover, the maps are usually hosted by the system provider. Since the systems are usually commercial, their free usage is usually limited by the number of map accesses.
- Programming libraries - e. g., Leaflet, Mapbox GL, OpenLayers, Google Maps Platform. Those libraries allow users to create various types of thematic maps without any radical limitation. They, however, require the users to have decent coding knowledge and understand the principles of geo-visualization (such as knowledge of geospatial formats - e. g., GeoJSON). Those libraries usually do not offer a palette of thematical maps but allow to create them by using essential graphical elements (such as layers, points, polygons).
Both approaches have advantages and disadvantages. A prototype of a compromise solution-Geovisto library-was developed during the TACR project (19/20) in cooperation with the Brno University of Technology, Masaryk University, and the Flowmon company. The library is based on the Leaflet library. However, it tries to provide a higher level of abstraction by offering a set of thematic maps (e. g., cartogram, connection map). Those thematic maps can be configured either programmatically or by using controls known from popular authoring systems. The library targets programmers who want to use a ready solution, and they do not want to be limited by the need to deeply study the framework and the requirements of proprietary authoring systems.
The doctoral thesis's goal will be to study existing approaches to geospatial data visualization and look for new ones. Particularly:
- analyze the process of the design and creation of thematic maps and propose its simplification
- design new algorithms for a pleasing arrangement of graphs in the map of (nodes and connections topologies) - e. g., based on edge bundling algorithm
- apply advanced animations to present geospatial data in time (geo-temporal data analysis)
- design performance optimizations for processing of large amounts of geospatial data records (big data processing)
- use geo-visualization and thematic maps for educational purposes (gamification)
- or according to the current interest of the student
The results will be implemented and practically validated using the Geovisto library, or the other possibilities of validation will be researched.