theto 0.2.1

Workflow automation for exploring location data

Theto: workflow automation for exploring location data

Any visualization requires a large number of trivial decisions that, when added up, result in a non-trivial impact on the quality and usability of the visualization. This is especially true of geospatial visualization, where data needs to transformed fit a particular map projection or a particular map tile service needs to be configured to contextualize the plot. Theto abstracts a lot of this overhead so more time can be devoted to looking at and understanding the data. A Theto instance allows users to:

1. Store api keys, palettes, and other static resources repeatedly needed throughout a typical visualization pipeline.
2. Load data sources, format coordinate data to stage it for all foreseeable downstream needs, and append metadata.
3. Add widgets to interactively filter what the final visualization shows, and infer the appropriate parameters for those widgets based on the source data.
4. Determine plot bounds, size, map zoom level, and other parameters in ways that accommodate the data.
5. Add layers of visualization, including tooltips and other visual aids, including connections between data points.
6. Render the plot, either in the notebook or by saving to file, optionally appending an interactive legend.

"Theto" is a transliteration of the Greek word Θέτω, which means "I place" or "I situate", or simply "I put". That's what the tool does: it puts geospatial data where it needs to be so users can spend their time looking at and making sense of what's there.

A Jupyter notebook demonstrating a lot of Theto's functionality can be found here:

https://nbviewer.jupyter.org/github/Valassis-Digital-Media/theto/blob/master/theto_demo_notebook.ipynb

Installation

To install from PyPI:

pip install theto

conda installation coming soon.

Supported data representations

Data can be loaded in a variety of formats (geohashes, WKT, shapely objects, GeoJSON, or coordinate pairs). The tool will automatically detect the format and process it appropriately. Likewise, any input can be rendered as the original shape itself (a polygon) or as the centroid of the shape (a point). Polygons are rendered using Bokeh's MultiPolygons glyph. Points can be rendered using any of Bokeh's marker glyphs.

Limitations

Theto is designed for interactive exploration, and is therefore appropriate for small-to-medium sized data. A very rough benchmark indicated that it takes about 5 seconds to plot every 50,000 points (a polygon might contain very many individual points) in a Jupyer notebook, and the notebook freezes at around 250,000 points. Outputting to file and viewing in a separate browser window should allow up to around 1 million points. For larger visualization, see http://datashader.org/.

Contributing

We welcome issues and pull requests that help improve the variety of data sources and plot elements Theto supports, its usability, and its re-usability within other tools.

License

Copyright (c) 2019 Valassis Digital under the terms of the BSD 3-Clause license