data visualisation

Data visualization is a discipline created by John Tukey.

Principles

• Choose the right chart: For instance, bar are good to show frequencies, absolute quantities. Lines are good to show variations over time, it can also show relative quantities, as a zoom to highlight some difference.

• Avoid misunderstanding: For instance, the bar must be proportional to what it represents.

• Choose the right colour scheme: if colours are showing important differences, so they must be easily differentiated. It is also good to avoid too much

• Remove unnecessary information: avoid any label and extra data. Less is more!

• Have in mind the story the chart is supposed to tell: at the end, it is expected to tell some story giving to the user some information. It is important to make sure the chart is delivering it.

Navigating

• This topic is related to many disciplines among them: machine learning, writing and math.

Sources

• Drawing from data presents some challenges on data visualization and some possible solutions, for instance, how to filter and aggregate data and how to make a pairwise distance matrix.

• The Beauty of Data Visualisation briefly talks about the importance of this discipline giving a few examples.

• Data is ugly and data is beautiful are reddit communities (subreddits) for discussing data visualisation

• Some beautiful customized palettes: 20 unique and memorable brand color palettes to inspire you and branition

• Seaborn framework documentation also brings useful information about palettes.

• Covid-19 dashboards in Java data visualization using vaadin an open-source framework for development of progressive web applications.

• The 25 Best Data Visualizations of 2018

• WHAT CHARTS DO

• A diagram type for a niche data use case