Code librarySample
Vega-Lite
University of Washington Interactive Data Lab
A declarative grammar of interactive graphics: describe a chart in JSON, or in Python through Altair, and the renderer does the rest.
Licence
Open source
Built by
Developers
Languages
JSON, Python (Altair)
Runs on
Web browser, Jupyter notebooks
Scores
Out of 5. The tick marks the average for code libraries in the library.
- Ease of use
- 3
- Visual control
- 4
- Interactivity
- 4
- Sharing
- 3
- Governance
- 1
- Low cost
- 5
- Data scale
- 2
Strengths
- Short, declarative chart specifications
- Interaction (selections) is part of the grammar
- Specs are portable and easy to generate from code
Limitations
- Less control than D3 for unusual visuals
- In-browser rendering struggles with very large datasets
- A smaller ecosystem
Best for
- Reproducible charts in notebooks
- Charts generated from code
Other code libraries
Matplotlib
Open-source community (NumFOCUS)
The foundational Python plotting library: low-level control over every element of a figure, for static, publication-quality output.
Open sourcePython
seaborn
Open-source community
A statistical plotting library on top of Matplotlib: one call for distributions, regressions and grouped comparisons from a pandas DataFrame.
Open sourcePython
Plotly
Plotly
Interactive, web-based charts from Python, R or JavaScript, with Dash for turning them into complete analytical apps.
Open sourcePythonR