Archives for posts with tag: ifgi

Today I have the chance to present at the first developer forum in the context of the ESRI Development Center (EDC) programme. The workshop has the topic “Space & Time“, and brings together developers and users from both companies/public agencies and universities to talk with each other about ongoing development for spatio-temporal analysis. A focus is the exchange about how to integrate different technologies.

On the first day there were presentations about concepts of space and time in GIS, applications and different research projects. It was really interesting to bring researchers and users together, and we had good discussion, imho also because all people have a developer background in at least ArcGIS or R. We also had a little bar camp and discussed about integration of R and ArcGIS, which I find has a high potential not only comparing analysis features, but from the possibilities that ArcGIS could offer in the dissemination are – imagine uploading an R script (maybe wrapped in Python/ArcPy code) to an ArcGIS Server and immediately having a WMS with you analysis output… We concluded that a survey has to be done about the users benefits for both communities.

On the second day there is a hands-on workshop with R on the one hand and ArcGIS on the other – and this is where I come in and present sos4R as a tool to import spatio-temporal data in R. Here I will actually not give a long presentation but prepared an R script including. Building upon that Edzer Pebesma will do some spatio-temporal analysis with that. And there you’ll not only get nice images but also a introduction to the package spacetime.

You can find the script files on the EDC page of the Open Geostatistics community.

Within the UncertWeb project the ifgi provides a SOS with parts of the AirBase dataset provided by the EEA. Release 0.1-15 contains a great demo on what you can do with that data: How to build up a spatial interpolation, in this case inverse distance weighting (IDW), on top of a GetObservation result. This comprises some data formatting and spatial projection.

Thanks go to Edzer Pebesma for contributing the interpolation examples using the packages gstat and spacetime and pointing out bugs in the package!

Some plots generated by the demo are:

  1. overview plot of the available offerings,
  2. a histogram of all downloaded NO2 values,
  3. a bubble plot of the data for Germany,
  4. the stations in Germany in an UTM projection,
  5. two IDW interpolations, one aggregated over time and one for a specific time stamp, and last but not least
  6. a time series plot for a whole year for one station including a polynomically fitted regression line.

Go ahead and run the analysis yourself:


library("sos4R")

demo("airquality")

It will be a bit quiet around sos4R for the rest of the year, so I’d like to share some older work today, one of the first plots I made based on ifgi’s WeatherSOS.  It is now part of a demo of the package, so it’s very easy for you to produce it yourself, and the demo even contains another plot.

# load sos4R package
library("sos4R")

# start the demo
demo("weathersos")

# press 'Enter' to go through the available plots

The first plot contains two lines showing the temperature in Münster and Kärnten for a week in August 2009:

WeatherSOS Temperatur in Münster and Kärnten

The second plot shows the station just in Münster for the month of September in 2010 together with a (polynomial) fitted line:

WeatherSOS Demo Plot 2, Temperatur in Münster with regression lineThere are more demos, which are currently the best way to get to know the package. Check them out!