
As a Geospatial data scientist, 2019 brought some new tools that made my life easier. In this post, I am sharing the best of these new additions in the Python ecosystem and some resources to get you started. You will find tools that accelerate your Geospatial Data Science pipelines using GPU, advanced Geospatial Visualization tools and some simple, useful Geoprocessing tools. I hope you will find one or two from the list that can help you.
1. cuSpatial: GPU-Accelerated Spatial and Trajectory Data Management and Analytics Library
It is part of open-sourced libraries with GPU accelerated data science pipelines entirely carried out in the GPU. It was one of the continually developing libraries in 2019 and part of these ecosystems, cuSpatial provides accelerated tools for Geospatial data science processes. A project showing the capability of the GPU accelerated Geospatial data science can be found here: Where should I walk? And the releasing blog.
So far, cuSpatial supports the following features with other features planned in the future:
- Spatial window query
- Point-in-polygon test
- Haversine distance
- Hausdorff distance
- Deriving trajectories from point location data
- Computing distance/speed of trajectories
- Computing spatial bounding boxes of trajectories
2. KeplerGL Jupyter Notebook
KeplerGL Jupyter is both convenient and remarkable tool for Geospatial data visualisation inside Jupyter notebooks. This library provided Uber’s advanced Geovisualization tool for Big and was released in June 2019. It allows users to plot maps inside Jupyter Notebook easily and also comes with the Interactive User Interface (UI).
If you want to get started with this tool, here is a step by step guide on how to use in Jupyter notebook.
Kepler.GL & Jupyter Notebooks: Geospatial Data Visualization with Uber’s opensource Kepler.GL
3. Pysal 2.0: Python Spatial Analysis Library
Pysal is primarily for doing spatial statistics in python and with the release of 2.0 brought it with an improved level of integration with other Geospatial libraries like Geopandas. Pysal now offers a collection of tools for geographic data science packages in Python, including Exploratory Spatial Data Analysis (ESDA).
Here is a guide on how to perform ESDA with Pysal.
4. GeoFeather
Geofeather was small but a great addition to the Python ecosystem. This library brings a faster file-based format for storing geometries with Geopandas, just like a feather is to python. With a simple experiment, I found Geofeather is super fast. I had to wait 15 minutes to write shapefile with Geopandas. It only takes 25 seconds to write that same file in Feather with Geofeather.

5. PlotlyExpress
PlotlyExpress, although not specifically for Geospatial data visualisation, it brings an easy and intuitive API for Geographic data visualisation with less code. Chances are you have seen PlotlyExpress chart or map already in other articles here in Towards Data Science, as this was immediately adopted in the data science community.
6. EarthPy
EarthPy promises to integrate both Vector and Raster data processing into one python package. It is built on top of Geopandas and Rasterio. For current spatial analysis, we deal with multiple libraries. Combining some of these functionalities holds some value in many cases. Some resources and examples are available here.
7. PyGeos
PyGeos was a new addition to the Geospatial Python Ecosystem. It provides vectorised functions to work with arrays of geometries (Numpy), giving better performance (than Shapely) and convenience for such use-cases.

8. Momepy: Urban Morphology Measuring Toolkit
MomePy was released late 2019, and it is an excellent addition for the Urban planners and Geospatial data scientists, enabling them to perform advanced quantitative analysis of urban morphology. The library comes with extensive examples using Jupyter notebooks
Other exciting additions not specific for python and Geospatial data science include Streamlit (for web applications), BigEarthNet (Geospatial Big dataset), #Aerialod (3D Geospatial Visualization) and BlazingSQL (GPU database).
Conclusion
Another year has passed with significant and meaningful open-source project contributions. These are some of my personal favourite tools in 2019, let me know some of your favourite Python Geospatial data science libraries in 2019.