Telling Stories with Data

by Professor Rhys C Jones, University of Surrey (FHMS)

Dr. Rhys was a Mentor in Residence in September 2021, and shared his expertise in teaching and learning math skills for researchers. Rhys created a hands-on workbook to help those of us who need a refresher for the math we’ve forgotten, or who need to develop new skills: Essential Math Skills for Exploring Social Data. Use the code SAGE30 for a discount when you order his book from SAGE.

The communication of information is ubiquitous; however, knowing the most effective way to present and explain data can be a challenge.

Channelling data into a story well, can leave a lasting impression on readers, convincing them of a particular argument or stance. Data stories can be overlooked by authors, which includes the careful selection and presentation of data into a coherent narrative. This blog will provide some tips and advice for presenting data stories, including the importance of knowing the intended audience. The advantages of using descriptive data, tables and graphs are also discussed.

Creating a data story requires the use of multiple skills to ensure there are harmonious connections between your main body of text, and the graphs and tables used to draw the reader to specific points of interest, central to the big ideas in the data story. Here are some tips for bringing data stories to life. The language you use will depend greatly on the intended audience. For example, if it’s an academic piece of work, more advanced or technical language may be required.

Use:

1.     Language that people understand

2.     Short sentences, short paragraphs

3.     Simple language: “Get,” not “acquire.” “About,” not “approximately.” “Same,” not “identical”

4.     Bulleted lists for easy scanning

5.     Numbers in a consistent fashion: For example, choose 40 or forty, and stick with the choice made

6.     Rounded numbers (by rounding both long decimals and big numbers)

Avoid:

1. “Elevator statistics”: This went up, this went down, this went up

2. Jargon and technical terms

3. Using large numbers, which can be difficult to grasp. Use the words millions, billions or trillions. Instead of 16, 956, 990 write “about 17 million.”

4. Using percentages for small numbers (i.e., sample or population sizes that are less than 50)

5. All capital letters and all italics: Mixed upper and lower case is easier to read

Data are often presented in tables to reveal interesting patterns or relationships.

Being able to read tables, spot patterns of interest, and then communicate those findings are extremely valuable skills applicable across a range of disciplines. Tables can include data involving percentages, proportions or ratios, or the actual values from the variables themselves (e.g., distance between participants homes and school). If percentages are used, it’s a good idea to include the sample or population sizes in the table. Also, if numeric data are to be included, think about the level of accuracy that should be used. For example, reporting the number of children adopted in a certain region in 1932 could include the use of whole numbers, or mean values that could include decimal places.

Good tables should complement and be connected to text. They should present numbers in a concise and organized fashion to support the main points. Tables help minimize numbers in the data story (i.e., the main body of text). Tables are generally small and shouldn’t be too large or complex. One decimal place will be adequate for most data. In specific cases, however, two or more decimal places may be required to draw attention to subtle differences in the data. Tables should also contain a title that is clear and concise. 

Displaying data in graphical format is also a common way of presenting data, found in many disciplines. Whenever presenting a graph, these questions can help to ascertain the point of them, and the intended messages to be conveyed:

• What are the main features of the graph?

• What other details are useful for understanding the variable/s?

An effective graph has a clear, visual message, with an informative and concise heading.

Good statistical graphics should:

1.     Present logical visual patterns that are not too busy, or contain too much information

2.     Have the appropriate scale on the axis. For example, reporting percentage increases over time can be big or small values. Making the scale very small (i.e., percentage increments of 0.1%) would make the percentage increase look a lot bigger than it is

Achieve clarity in graphs by:

1.     Using data values on a graph only if they don’t interfere with the reader’s ability to see patterns of interest in the data (i.e., the message that you really want to convey)

2.     Making all text on the graph easy to understand

3.     Not using abbreviations

4.     Avoiding acronyms

5.      Avoiding legends except on maps