There will be a short historical overview of data visualization with some examples and trends.
A wide selection of platforms are readily available on the market, of which most are paid services. These platforms offer a large variety of tools to store, explore and present data. They save a lot of time on learning and programming side and are suitable for companies, where data flows through on already tested pipelines and the focus is on the results. On the other side of the spectrum are Jupyter notebooks and Rstudio (or Mathematica), which are suitable for the development of both new data exploratory tools and new type of presentations.
There is always a trade off between quick plotting, which are standard and not really customizable, and a highly customizable environment, where programming and an understanding of the underlaying language is necessary.
This lecture's main focus are on freely available python packages, that utilize the capabilites of browsers. We will cover many key tools hosted at pyviz.org (holoviews, panel, geoviews etc.) and related modules (bokeh, param etc.). We should have a bit of insight into the technology of these modules and understand their limits and capabilities. The task will be to create a static, interactive and a hosted report/presentation.
http://www.haralick.org/DV/Handbook_of_Data_Visualization.pdf - A lengthy handbook from the prejupyter era
Preparing graphics is a high level task
https://www.slideshare.net/thompsonkaren/data-visualization-theory-59604862
http://datavis.ca/milestones/index.php
It used to be static. These displays should be of high quality and include complete definitions and explanations of the variables shown and of the form of the graphic.
Some rules for preparing: http://ling.uni-konstanz.de/pages/home/butt/main/material/lingvis/VisTheory_v1_online.pdf
Fast, dynamic, not necessarily precise, but adds new perspectives for the viewer, that helps to grasp the meaning of the results.
Checkout the gapminder folder, which contains a Bokeh application!