By Diego Taborga from the Knowledge and Learning Sector at the Inter-American Development Bank
Who doesn’t like a good story? I recently read an article from Paul Zak for the Harvard Business Review, called “Why Your Brain Loves Good Storytelling”. In the article he describes how character driven stories were able to release higher doses of a neurochemical called Oxytocin. This chemical “is produced when we are trusted or shown a kindness, and it motivates cooperation with others”.
The power of storytelling lies in its ability to change people´s behavior or knowledge. Consequently, it has a lot of potential for getting out the messages and lessons that our data are telling us. Data stories are increasingly being used to engage audiences with the most important messages that come out of rigorous data work. Of course, a relevant question then becomes, once I have analyzed data and identified the key conclusions that we can draw from it, how can I use storytelling to present the ideas. There is no doubt on the power of data storytelling but you must always be clear about your message and choose whether you will produce a video, an infographic, an illustration or any other format according to the what you want to convey.
The following steps will guide you in the process of developing a data story.
Define a call for action
This is probably one of the most challenging steps but yet the most necessary. What is the message you want to transmit? What do you want your readers (or listeners) to share or to do after hearing the story? We use data stories to transform something abstract or complex into something that people will easily understand, contextualize and remember (“anecdote trumps data”).
The identification of the call for action allows us to focus the design and narrative of our story in a way that will increase the likelihood that persons share the story, commit to specific actions and/or make desired changes.
Choose the audience
Pixar, the well-known animation studio, recommends that Rule 2 of the 22 rules for storytelling is to know who the audience for our story will be and what is interesting for them: “You gotta keep in mind what’s interesting to you as an audience, not what’s fun to do as a writer. They can be very different.” Consider who will be interested in the message that your data tells and to whom are you issuing the call for action. Write the story to them.
Imagine that your chosen audience had an opportunity to dive deep into the data; what are the questions they may come up with? What type of decision-making they will be facing and what information do they need for that? Make sure you include their perspective into your story.
It is important that your audience can review the data if some questions come up or if they do not completely agree with your story. During the collection, processing and analysis of data, you have made a lot of decisions regarding what to report and how to record it; you interpreted certain variables in specific ways. It can be overwhelming to remember all the parameters and decisions that you applied. So, make sure to document them and share relevant information and caveats with your audience.
Always keep it simple
When you present your data, keep your audience in mind. You should not mention terms such as standard deviation or confidence interval if your audience is not statistically saavy. Here, Jeremy Taylor goes over a few steps that may help you throughout this process. For instance: starting with a look at descriptive statistics first, trimming your data beforehand in order to focus on the analysis, or even accepting that you may end up with no significant data. For deeper reading on how to present statistical data, you can also read the U.N. manual on “Making Data Miningful.”
Use visualizations to complement the narrative
Visualizations enhance the story that you are telling, but the choice of what and how to visualize is important. Meg Cannistra, in her Ceros blog, suggests that choices should be based on the type of data relationship that you are telling about. Choosing how you will tell your story can be overwhelming. You must keep in mind what is the purpose of your message. For instance, here is an article on different types of visualizations and how to choose according to you message. Moreover, in “Narrative Visualization: Telling Stories with Data“, Stanford researchers discuss author-driven versus reader-driven storytelling; do you want to set a rigid message or do you want to promote audience interaction with the data?
Tell an honest story
In general, we as humans tend to make the mistake of seeing what we want to believe. If your hypothesis when interpreting your data was not clear, you can sometimes “find” things in the data that are not valid but support our personal views. So make sure you are telling an honest story that is consistent with the data, regardless of the message you want to transmit or the decision you may want to influence. For further information, Steve Cooper has 7 Tips When Analyzing your Analytics. For instance: remembering what the context is, compiling “items of interest,” looking for shifts, etc.
Have a data story to share? Include it in the comments below!
Has a degree in Political Science and Psychology from the University of Kansas.
Diego es un ciudadano americano y boliviano que trabaja como consultor en el Banco Interamericano de Desarrollo para del Sector de Conocimiento y Aprendizaje. Actualmente se encarga de dar apoyo al proceso de diseño, implementación y evaluación de los planes de conocimiento y aprendizaje desarrollados por departamentos internos, con el objetivo de desarrollar una mayor eficiencia y efectividad operativa. Además de apoyar el desarrollo de soluciones integradas para las necesidades de los clientes y establecer asociaciones para ofrecer servicios de valor añadido, su área de interés personal se enfoca en cómo mejorar la implementación de estas actividades mediante el uso de datos.
Diego cuenta con una Licenciatura en Ciencias Políticas y Psicología de la Universidad de Kansas.