TRACLUS Trajectory Clustering in C++ and R

TRACLUS is a seminal approach to trajectory clustering that combines the flexibility of density-based clustering with the spatial complexities of trajectories. In a certain sense, TRACLUS is a framework consisting of first splitting trajectories into sensibly local micro-segments, then using a variant of DBSCAN to cluster these segments, followed by grouping operations making sure that clusters of trajectory segments come from clusters of trajectories and are not artifacts of the splitting strategy. A free implementation in C++ with bindings for the R programming language is provided.

Read more on this…


GISCUP’15: Routing Under Polygonal Constraints

The award-winning implementation of routing in street networks under dynamic polygonal constraints and its source code. It is a careful implementation including data loading, reprojection, and visualization, in-memory spatial indexing using a variant of the R*-tree for the GUI (mouse clicking), Dijkstra, A*, and an optimized version of the ALT algorithm that was able to win the challenge due to the fact that the precomputations are compatible with increasing edge weights and that the precomputation time is surprisingly low. It is not the right approach to solve dynamic routing in continental-scale problems. For this direction, look up personalized contraction hierarchies...

Read more on this…


DSLAB: Portable, Interactive platform for Data Science Demonstrators using OpenGL

A software package to create GUI-like applications with console in and output going to a side bar and OpenGL being used to show the main information. Specialty is that it can be cross-compiled from Linux to Windows using MXE.

Read more on this…