There are many ways to visualise data but here are some explainers to walk you through the maze of its vocabulary and help you decide how to tell your interactive story.
Visualising complex statistical and numerical data in a comprehensible visual form is a carefully crafted task.
If the news story is trying to illustrate comparisons, to show differences or similarities, those most synonymous with axes like a Bar Chart or Line Graph are often used. But Bubble Charts, Scatterplots along with Radial Column or Radial Bar Charts can also be explored.
A few of the above, especially some line graphs, columns and scatterplots are often used by newspapers such as The Financial Times. These can be to discuss topics of correlation like life expectancy and incomes or inflation and unemployment or simply just to show a pattern over time. However, there is a need to explain to the readers the relationship between the variables as unintentionally some might assume that one causes the other.
One drawback of bubble charts, however, is its limited ability to show large data beacuse too many bubbles can make the chart difficult to read, as demonstrated below. Although incorporating interactive elements such as hovering or clicking the bubbles can improve this by revealing unseen information.
To pinpoint geographical location or distribution in an article, spatial charts are increasingly being used to draw attention of readers. Most recently they have been used to show election results and their variations across a country or region. Some examples of their use can be found in reports on the German national elections, UK’s referendum on Brexit and the presidential election in the United States.
Spatial charts are also seen as useful tools to show population density, draw attention to the impact of natural disasters or illustrate areas at risk in future by such events. The knowledge of precise geographical coordinates is necessary for a visualisation like this.
Basic Choropleth, Heat and Contour are some of the maps used to visualise data as well, but Equalised Cartogram and Scaled Cartogram are also more engaging ways to explain.
If you are starting off with basic Choropleth Maps, its use of geographical areas with the help of colour and shades to show variation and patterns may appear exciting. However, sometimes one of the drawbacks of colour makes it difficult for the reader to accurately compare values and at times larger regions also may appear unnecessarily more important.
It is also essential to normalise values when, say making a map based on population (for example calculate population per square kilometre) rather than place raw data values on to it, as this can be misleading. Below is a good example of Britain’s EU Referendum result based on normalised population value in a Choropleth Map.
Visualisation tools to generate some geo-spatial graphics: Basic Choropleth: Datawarpper and Kartograph (code)
To demonstrate the importance of movement over time in a report some of the graphics called for are Connection Maps, Flow Maps, Networks or Chords. Overall these are helpful to communicate movement intensity between geographical areas or conditions. They have been used to explain the movement of refugees, information and trade. Financial journalists also call upon them to describe activities of markets and funds.
Connection Maps are an engaging way to demonstrate geographical links and routes or show how concentrated connections are on a map. Whereas Flow Maps are often used to show the movement of people, animals and birds which helps readers understand the geographical distribution of the subject being discussed.
Although a large amount of data shown on Flow Maps can appear visually cluttered, this can be reduced by merging and bundling.
Chords according to Financial Times Visual Vocabulary is a complex yet powerful diagram. It uses it to explain ‘two-way flows and net winners in a matrix’. Networks, the same vocabulary describes has the strength to ‘demonstrate inter-connectedness’ of varying types.
The Connection Map below illustrates data of the routes migrants have taken to reach safety due to recent conflict or regional instability.
When a story calls for showing data over time often Histograms or Line Graphs can be used but big dataset can utilise Spiral Plots to their advantage as it can hold trends over a large period of time and colours can be used to show comparisons. The plot can use lines, bars or points along a spiral.
Stream Graphs are also a visualising method using flowing organic shapes resembling a river stream which can show trends over time. It is also a variation of Stacked Area Graph which like the Spiral Plot can handle large datasets to show patterns over time and categories.
One of the disadvantages, however, is its visual legibility when illustrating very large datasets as smaller categories go almost unnoticed but incorporating interactive elements in that Stream Graph can resolve this.
Change over time can also be shown using Circle, Sequential or Scaled timelines. These can negotiate time across decades or centuries, for example, to show natural disasters like earthquakes by continents or conflicts over time.
Visualisation tools to generate some data over time: Spirial Plot: Arpit Narechania’s Block (d3 code) Stream Graph: JSFiddle (code) and data360r Timeline: Timeline.js
To exemplify distribution of wealth or population across age and gender charts such as Density Plots, Population Pyramid, Pictogram Chart, Histograms and Violin Plots can be applied.
Density Plots are said to have an advantage over Histograms as they are not affected by the number of bins (a term used for a bar in a Histogram) so can still produce visually effective data for an audience.
Icons are used in Pictogram Charts which is effective when using small dataset. Each icon can be a unit, or any number of units and each category is compared across columns or rows. Pictograms are also successful in overcoming language barrier but avoid using them for large datasets and displaying partial icons as this might confuse an audience.
Violin Plots are most effective when describing data which cannot be summarised in simple averages. Hence it is useful for conveying complex distribution data and its probability density.
Watch Hans Rosling's Yardstick of Wealth - Don't Panic - The Truth About Population - BBC Two
Visualisation tools to generate some distribution datasets: Density Plots: R Graph Gallery Pictogram: jChart Violin Plot: R Graph Gallery