Training, Tutorials and Templates
OxShef: Charts provides general advice on choosing the most appropriate visualisation for your data and to avoid making common mistakes or pitfalls in data viz. However, we do not provide interactive tools for matching your data to charts (or vice versa).
There are many tools (Excel, Google Sheets, Plotly, Tableau) which automatically suggest charts based on the properties of your dataset. This is possible thanks to excellent research into visual perception theory and machine vision from back in the 1980s through to today.
In fact, some tools like ReVision are even capable of taking existing charts and re-designing them to be more accessible and easy to read by humans. The example here shows a number of badly designed pie charts that have been converted into much easier to understand barcharts.
OxShef: Charts maintains a collection of tools and resources to assist you in developing and designing effective visualisations, split into the following three categories:
Choosing a dataviz
Some charts are more suitable for specific purposes than others. In some cases these resources can automate the chart selection process, or at least help exclude some chart options.
Designing good dataviz
There are many best practices for dataviz which radically improve the legibility of a chart, for instance: horizontal barcharts with bars arranged from longest to shortest are significantly easier to interpet than unordered vertical barcharts.
Adding interactivity
Sadly, there’s little general advice for effectively adding interactivity to charts. However, two extremely useful rules of thumb are “Make the least interactive thing that works” (@BrianBoyer) and Ben Shneiderman’s mantra Overview first, zoom and filter, then details-on-demand.
Ben Shneiderman's mantra | Ben Shneiderman's mantra is one of the few general purpose pieces of advice for adding interactivity to dataviz:
Overview first, zoom and filter, then details-on-demand | Research paper | |
Chart Maker (Visualising Data) | Andy Kirk designed and maintains the Chart Maker as a tool for selecting data visualisation tools based on specific charts. Andy has put a lot of effort into designing minimal explanations of each chart, this is a great reference tool when selecting both charts and dataviz tools. | Interactive tool for comparing different charts | |
Eager Eyes | Eager Eyes provides long-form articles and investigations into a wide variety of data visualisation topics and best practices. This website is not only useful for designing good dataviz, but also in choosing which chart to use in the first place. | Robert Kosara’s website and blog dedicated to long-form articles on data viz. | |
Financial Times Visual Vocabulary | The Financial Times has invested significantly in their data visualisation toolkit, part of which is a "visual vocabulary" that they use to help automatically choose the most appropriate dataviz for each dataset. In the future the FT are planning on open sourcing their tools based on this visual vocabularly, for the time being this remains a useful cheat sheet. | Poster presenting the Financial Times' taxonomy of dataviz types. | |
Why We Are Doing Fewer Interactives | Archie Tse from the New York Times spoke at the 26th Infographics World Summit in 2016 about how their team has reduced the number of interactives they've developed in response to user testing. Archie's overall advice can be summed up as "if anything other than scrolling is required, make something spectacular happen". | Conference presentation |
