Skip to content
CUDA-accelerated GIS and spatiotemporal algorithms
Python Cuda C++ CMake Shell
Branch: master
Clone or download
Pull request Compare This branch is 4 commits ahead, 63 commits behind branch-0.11.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Type Name Latest commit message Commit time
Failed to load latest commit information.


GPU-Accelerated Spatial and Trajectory Data Management and Analytics Library

NOTE: cuSpatial depends on cuDF and RMM from RAPIDS.

Implemented operations:

cuSpatial supports the following operations on spatial and trajectory data:

  1. Spatial window query
  2. Point-in-polygon test
  3. Haversine distance
  4. Hausdorff distance
  5. Deriving trajectories from point location data
  6. Computing distance/speed of trajectories
  7. Computing spatial bounding boxes of trajectories

Future support is planned for the following operations.

  1. Temporal window query
  2. Temporal point query (year+month+day+hour+minute+second+millisecond)
  3. Point-to-polyline nearest neighbor distance
  4. Grid-based indexing for points and polygons
  5. Quadtree-based indexing for large-scale point data
  6. R-Tree-based indexing for Polygons/Polylines

Install from Conda

To install via conda: conda install -c rapidsai-nightly cuspatial

Install from Source

To build and install cuSpatial from source:

Install dependencies

Currently, building cuSpatial requires a source installation of cuDF. Install cuDF by following the instructions

The rest of steps assume the environment variable CUDF_HOME points to the root directory of your clone of the cuDF repo, and that the cudf_dev Anaconda environment created in step 3 is active.

Clone, build and install cuSpatial

  1. export CUSPATIAL_HOME=$(pwd)/cuspatial
  2. clone the cuSpatial repo
  1. Compile and install C++ backend
mkdir build
cd build
make # (or make -j [n])
make install

cuSpatial should now be installed at $CONDA_PREFIX. The cuspatial include path is $CONDA_PREFIX/include/cuspatial/ and the library path is $CONDA_PREFIX/lib/

  1. Compile and install cuSpatial Python wrapper and run Python test code
cd $CUSPATIAL_HOME/python/cuspatial
python build_ext --inplace
python install
  1. Run python test code

First, add the cuSpatial Python API path to PYTHONPATH (there are tools under tests subdir): export PYTHONPATH=$CUSPATIAL_HOME/python/cuspatial

Some tests using toy data can be run directly, e.g.,

python  $CUSPATIAL_HOME/python/cuspatial/cuspatial/tests/

However, many test code uses real data from an ITS (Intelligent Transportation System) application. You will need to follow instructions at data/ to generate data for these test code. Alternatively, you can download the preprocessed data ("locust.", "its_4326_roi.", "itsroi.ply" and "its_camera_2.csv") from here. Extract the files and put them directly under $CUSPATIAL_HOME/data for quick demos. A brief description of these data files and their semantic roles in the ITS application can be found here TODO THIS IS MISSING

After data are downloaded and/or pre-processed, you can run the python test code:

python  $CUSPATIAL_HOME/python/cuspatial/cuspatial/tests/
python  $CUSPATIAL_HOME/python/cuspatial/cuspatial/tests/
python  $CUSPATIAL_HOME/python/cuspatial/cuspatial/tests/

NOTE: Currently, cuSpatial supports reading point/polyine/polygon data using Structure of Array (SoA) format (more readers are being developed). Alternatively, python users can read any point/polyine/polygon data using existing python packages, e.g., Shapely, to generate numpy arrays and feed them to cuSpatial python APIs.

You can’t perform that action at this time.