A lot has changed in the last 2 decades since the paper on “Spatial Statistical Analysis and Geographic Information Systems” was published by Anselin and Getis. Today, the central focus of GIS is on spatial analysis and the rich set of statistical tools to perform the analysis. Today the GIS database and analysis tools are not looked upon as different software. Spatial analysis is fully integrated in GIS softwares like ArcGIS and QGIS. Furthermore, for very specialized applications, the modular or the loosely coupled approach is often employed. Software like CrimeStat uses data in established GIS Sofware format, perform analysis on them and produce results for use in GIS softwares.
When it comes to the nature of spatial data, two data models have been widely accepted namely Object based model and field/raster based model. Extensive set of analysis tools have been developed for each of them. Data heterogeneity and relation between the objects are also taken into account by slight improvements over these two models.
Exploratory Data Analysis and model driven analysis have progressed hand in hand and complement each other. While new and innovative visualization and exploration tools help in understanding the data and the problem better. Software has evolved over time to perform complex non-linear estimations required for model based analysis.
However, Statistics and GIS is an ever evolving field and newer methodologies and techniques are developed everyday which pushes the boundary of cutting edge research further and further. Newer challenges in statistical analysis include handling Big Data and community generated spatial information. How these new challenges evolve will be very interesting to observe.