Content uploaded by Maja Kalinic
Author content
All content in this area was uploaded by Maja Kalinic on Oct 06, 2018
Content may be subject to copyright.
Combining Scripting Languages and Desktop GIS for
Spatial Data Science
Maja Kalinic*, Jan Wilkening** and Mathias Jahnke***
* Institute of Geography, Applied Geoinformatics, University of Augsburg, Augsburg, Germany
** Esri Deutschland GmbH, Kranzberg, Germany
*** Chair of Cartography, Department of Civil, Geo and Environmental Engineering, Technical University of Munich, Germany
Introduction
In spatial data science, there are two main approaches of analyzing and
visualizing data:
To use powerful open-source scripting languages via the command line
(such as Python and R) or
To use graphical user interfaces (toolbox wizards) in Desktop GIS.
We argue that spatial data scientists are most efficient when they combine
the best of both approaches (using tools such as R-ArcGIS Bridge, ArcPy,
Python API). We show how Python, R, Desktop GIS and their combination
can enhance spatial data science in a crime analysis case study.
Workflow
Selected References
Kalinic M (2017) Combining Open-Source Programming Languages with GIS for Spatial Data Science.
Master’s Thesis, Technical University of Munich.
LeBlanc, J., Elder, J., Bruce, C., Cook, T., Rodriguez, E., & Steiner, F. (2014) Definition and Types of Crime
Analysis. International Association of Crime Analysts, Overland Park, Kansas.
Contact:
Maja Kalinic
maja.kalinic@geo.uni-augsburg.de
Universität Augsburg, Institut für Geographie
Alter Postweg 118, 86159 Augsburg, Germany
Correlation matrix with correlation coefficients between variables
Python’s bar graph with crime counts per district
Conclusion
The combination of scripting languages and core Desktop GIS
functionality is effective and efficient for answering spatial data science
questions.
Python (pandas, numpy, matplotlib) and R (data.frame, dplyr, ggplot2,
corrplot) libraries are useful in identifying and visualizing spatiotemporal
relationships, crime trends and patterns.
ArcGIS Pro contains useful tools for crime analysis in its basic
functionality. In combination with R and Python, these tools facilitate
information extraction and visualization of criminal data.
Strategic
(Python) Intelligence
(R) Tactical
(R-ArcGIS) Predictive
policing
Analyzing and
visualizing the
spatial and
temporal
distribution of
crimes
Analysing crime
dependancy
towards time,
place and crime
type
Analyzing trends
and patterns,
potential
influencers of
unlowful
behavior
Forecasting and
prediciting crime
events
Crime analysis (LeBlanc et al., 2014):
Results
Emerging Hot Spot Analysis for Crimes in San Francisco with ArcGIS
Python’s line graph with crime counts per year
R‘s heat map of crime counts in relation to crime
types and months at which they were reported
Bubble chart scatterplot with crime
counts in relation to districts and days
at which they were reported