Algorithms for Map Generation and Spatial Data Visualization in LIFE

Lin, Ying-Chi
Abstract in English: 
The goal of this master thesis is to construct a software system, named the LIFE Spatial Data Visualization System (LIFE-SDVS), to automatically visualize the data obtained in the LIFE project spatially. LIFE stands for the Leipzig Research Centre for Civilization Diseases. It is part of the Medical Faculty of the University of Leipzig and conducts a large medical research project focusing on civilization diseases in the Leipzig population [86]. Currently, more than 20,000 participants have joined this population-based cohort study. The analyses in LIFE have been mostly limited to non-spatial aspects. To integrate geographical facet into the findings, a spatial visualization tool is necessary. Hence, LIFE-SDVS, an automatic map visualization tool wrapped in an interactive web interface, is constructed. LIFE-SDVS is conceptualized with a three-layered architecture: data source, functionalities and spatial visualization layers. The implementation of LIFE-SDVS was achieved by two software components: an independent, self-contained R package lifemap and the LIFE Shiny Application. The package lifemap enables the automatic spatial visualization of statistics on the map of Leipzig and to the extent of the authors knowledge, is the first R package to achieve boundary labeling for maps. The package lifemap also contains two self-developed algorithms. The Label Positioning Algorithm was constructed to find good positions within each region on a map for placing labels, statistical graphics and as starting points for boundary label leaders. The Label Alignment Algorithm solves the leader intersection problem of boundary labeling. However, to use the plotting functions in lifemap, the users need to have basic knowledge of R and it is a tedious job to manually input the argument values whenever changes on the maps are necessary. An interactive Shiny web application, the LIFE Shiny Application, is therefore built to create a user friendly data exploration and map generation tool. LIFE Shiny Application is capable of obtaining experimental data directly from the LIFE database at runtime. Additionally, a data preprocessing unit can transform the raw data into the format needed for spatial visualization. On the LIFE Shiny Application user interface, users can specify the data to display, including what data to be fetched from database and which part of the data shall be visualized, by using the filter functions provided. Many map features are also available to improve the aesthetic presentation of the maps. The resulting maps can also be downloaded for further usage in scientific publications or reports. Two use cases using LIFE hand grip strength and body mass index data demonstrate the functionalities of LIFESDVS. The current LIFE-SDVS sets a foundation for the spatial visualization of LIFE data. Suggestions on adding further functionalities into the future version are also provided.
masterarbeit_ying-chi_lin_final.pdf9.26 MB