by Stephen Childs

We've all seen beautiful data visualizations on the web and elsewhere. A good visualization can make a persuasive point or give new insights. How can you create beautiful and useful visualizations without too much effort? The secret behind each visualization is a set of well-organized data. In this case, we look at data from online surveys, which is not typically well-organized, and see how we can transform it and easily visualize it using the 'Grammer of Graphics' approach. The 'Grammer of Graphics' is a way of associating different data points and different aspects of a chart. You can provide the type of chart that you want, specify which data you want on the x and y axes and how you want to group you data and you will get a reasonable chart. As you do a data analysis, it is important to understand your data. A visualization tool that can quickly generate useful charts during analysis and also generate the finished charts for production is ideal. The altair package excels at both these jobs. The altair package allows you to create web ready visualizations using this approach using Python. This talk will demonstrate how altair can be used with survey data to get quick insights out of a survey,or any other data source. Altair is built upon Vega and Vega-Lite, which are JavaScript libraries. They work well with Jupyter notebooks and are useful for data exploration. The data behind the chart and the code for the chart itself is stored as JSON and can be included on any web page, so the visualizations are independent of Python.

About the Author

Talk Details