Grammar of graphics and its Interactive cousin Altair by Eberhard Hansis

362

There is no shortage of data visualization tools and libraries out there. Since they are all very different, they leave us with the question: what are the basic principles of good data visualization? This talk is an introduction to a particular way of building charts, by thinking in terms of a Grammar of Graphics.

A Grammar of Graphics defines simple concepts or “building blocks” from which a chart can be assembled. By combining a relatively small number of design concepts you can build an infinite number of (arbitrarily complex) charts. All the while, the vocabulary and syntax for describing the charts remains compact and consistent.

After explaining the concept I demo Altair, a Python charting library. Altair wraps the Vega-Lite standard, which implements an interactive grammar of graphics, designed mainly for the web. I show how to explore a sample dataset while building up an increasingly complex visualization, including some interactive elements.

Demo code and links to interesting material can be found at https://github.com/ehansis/altair_demo

Topic was presented at Data Visualization in Munich Meetup
https://www.meetup.com/Data-Visualization-Meetup-Munich/

Video

You can watch recording of the presentation in our YouTube channel:

Bio

Eberhard Hansis is a physicist by training, scientific programmer by profession and data geek by passion. He is the Co-Founder and CTO of Vebeto GmbH. Vebeto makes a universe of publicly available data accessible for the strategic needs of their customers. This involves the integration and joint analysis of a multitude of data sources, advanced analytics, and a visual presentation that is tailored to the target audience. Find more information about Vebeto at www.vebeto.de