Welcome to the
GIScience Lab @ TU Vienna!
Our Lab consists of a powerful cluster of machines that allow researchers and students to work with Big Data in the area of Geographic Information Science. The research performed in this lab includes topics such as Urban computing and Spatial Machine Learning.
In the era of Big Data, the availability of spatial information is also increasing at an unprecedented pace. In the scope of the Urban Computing domain, our research group aims at dealing with and exploiting Big Spatial Data for (Geo)Spatial analysis and applications. For example, in a recent research project, we conceived a new model to describe street intersections and realized a computational solution to analyze intersection data worldwide. The result is publicly available for use in research at our intersections website. This information can be used, for example, to make spatial predictions during navigation or to generate look-alike virtual cities.
As nowadays smartphones and other mobile devices equipped with GNSS technology are ubiquitous, it has never been easier and affordable to collect vast amounts of highly accurate trajectories of human beings. In combination with contextual spatial data of all kinds, human mobility can be scrutinized in depth, revealing peoples’ spatio temporal behavior. Based upon the tracking data it is possible to derive not only individual but also collective mobility patterns, which enable us to look further into the movements’ structures and regularities. The scope of human mobility covers besides others the prediction of traffic flows, the extraction of place types as well as compute the complexity of the given environment.
Geosemantics and Spatial Machine Learning
Geospatial Semantics describe how human categorize the environment. Machine Learning on the other hand provides powerful tools and methods for predicting but also describing complex data. The fusion of both domains enable to understand the way humans structure the environment. The obtained knowledge is paramount for different future and present research: creation of novel geospatial indicators, prediction of geospatial phenomena and processes, understanding of complex geospatial relationships as well as A.I. in geospatial research. In our research we utilize state of the art technologies such as Deep Learning and Linked Data, however, complement it with traditional methods of geospatial analysis.
The cadastre is core part of a land information system. It creates a spatial tesselation and creates unique identifiers for the areas. This allows matching information with spatial units. It's counterpart in the legal domain is the land register, which stores private rights, responsibilities, and restrictions for the areas defined in the cadastre. It is the basis for many governmental tasks including spatial planning and national statistics.
There are numerous reserch questions related to the cadastre. Some of them are:
- How to determine the spatial accuracy of a cadastre and how to communicated the result?
- How to support automatization in digital cadastzres?
- How to enrich the model, e.g., by connecting it to public law restrictions?
- How to deal with the changing reality after earthquakes or due to landslides?
- How to find strengths and weaknesses in existing systems and how to use the findings?