DATA FOR ADVOCACY

How can data be used to advocate social change?

Data by themselves are not particularly useful. To change behaviour, we need to extract information from data, filter this information so that it serves a purpose, and then visualise this information in order to turn it into meaningful, impactful stories.

WHAT ARE THE BEST WAYS OF REPRESENTING DATA?

Anyone can tell a story using data. One of the most important benefits of visualisation is that it allows us visual access to huge amounts of data in easily digestible visuals. Well-designed data graphics are usually the simplest and, at the same time, the most powerful.

TYPES OF DATA VISUALISATION

BAR CHARTS AND PIE CHARTS

Use them to communicate strong messages quickly

HISTOGRAMS

Much like a bar chart, use them to communicate strong messages. Histograms group numbers into ranges per bar.

LINE CHARTS

Good for displaying time series only.

MAPS

 There are different types of maps. Have a look here.

INFOGRAPHICS

A collection of eye-catching imagery, charts and minimal text to give a clear overview of a particular topic.

CREATIVE MEDIA

Artistic representations using a range of media

Organising your information

ORGANISING YOUR INFORMATION

Think about what the best organising principle for the information might be, rather than what is the most convenient.

You will need to:

  • Balance aesthetics and design with getting the right narrative and foregrounding the most important information

  • Use colour to make information ‘pop out’ and convey meaning (see the example below)

  • Make things immediately clear and accessible without dumbing down the issue

  • Remember that the success of advocacy and campaign communications can only be judged by how they affect an audience

  • Give your audience one clear task or message: for example, a call to change one small piece of behaviour, a volunteer sign-up option, or a donation link

EXAMPLES OF EFFECTIVE DATA VIZ

AERIAL SURVEILLANCE MAP

Journalist Peter Aldhous made a map for a BuzzFeed article about aerial surveillance of U.S. cities. The U.S. government claimed that these planes were following up on specific leads related to drug smuggling, human trafficking, and other criminal activities, though Buzzfeed reported critics’ views that the data showed a tendency to target particular communities by ethnicity and other characteristics. This example shows how effective visualisation can help to add value and meaning to a campaign designed to effect policy and practice.

PRESCRIPTION DATA

Mastadon C, in collaboration with the Open Data Institute, analysed 37 million rows of statin prescription data in different regions across England. They identified that if generic medication could be prescribed over branded medication, the NHS could potentially save £200m in a year.  The Prescribing Analytics website provides a visualisation of this finding using maps and clear labelling system of colour density.

Training your critical eye

Be aware that many data visualisations can be misleading – always take a critical view. Do the segments of a pie chart actually add up to 100%? Are the segments proportionate? Do the horizontal and vertical axes of a bar chart make sense? See viz.wtf for examples of bizarre, funny and downright wrong data visualisations.

Some data visualisations may be unintentionally misleading, but be alert to graphics which lead you to a conclusion that support specific agendas. For example, this graph was tweeted by Ann Coulter, a far-right US media pundit, to support the idea that the impact of COVID-19 had been overstated.

However the graph was misread as the journalist who explored the original figures, Andy Kiersz, wrote in his  article. For a start, the two visualisations show different age categories; they also compare two data sets without additional context – such as the fact that the US and South Korea have taken very different approaches to managing the coronavirus outbreak. Finally, the figures don’t support the assertion made in the tweet:  the death rate for South Koreans under 60 for COVID-19 is in fact over 7 times higher than the deaths in the US from flu in the under-65 range. This illustrates how bad visualisations are not only easy to misread, but can be actively misrepresentative.