Spatial Data Science integrates Geographic Information System (GIS) and Data Science. This article briefly describes Geospatial or Spatial Data Science in my point of view in simple way. Of course there are may other perspectives of Spatial Data Science. GIS contains 3 steps. They are data input, processing, and visualization/output.
The first step is data input, also called as data preparation, wrangling, management, storing, and many other relevant activities. The data input then will be processed or analyzed with methods. The methods can be statistics, spatial analysis, geospatial, Machine Learning, Deep Learning, and etc. The last step is to visualize, communicate, or present it effectively and efficiently. General GIS software, such as ArcGIS and QGIS, basically can do all the three steps above. By mastering fundamental GIS, many kinds of spatial work can be performed well. But, more knowledge and skill are needed for advanced spatial work. Basic GIS is not enough to handle large-scale spatial datasets.
Relational Database Management System (RDBMS), like PostgreSQL/PostGIS, MySQL, etc., is needed to perform bigger size data management and preparation. Spatial database is managed using certain extension, like PostGIS from PostgreSQL extension. R or Python also can do data management tasks, but limited. Basic GIS software, like ArGIS or GIS, is not enough to manage large-scale datasets.
Likewise, basic GIS software also has limited capability for data processing and analysis. R and Python are used to enhance the capability of analysing bigger size datasets, including programming the data automation, classification, prediction, and others. See how Python works for GIS analysis here.
Spatial data or information is finally visualized to communicate the result. Basic GIS software, again, has limited function to do this. Usually, maps can be created and printed on papers. But, in modern era, people need more than just a simple printed map, especially in looking for complex large-scale spatial and tabular data. Interactive map is required. Interactive map, WebGIS, or dashboard provides the user to zoom in, zoom out, pan, click, or hover the pointer while understanding the map. Another article here is my work of interactive dashboard visualizing spatial and tabular data. The dashboard is created using Shiny R, supported by web programming languages, like HTML, CSS, and Javascript.
Data Science and GIS 