This week we collaborated with Equal Measures 2030 and Tableau Public for International Women’s Day, using a dataset that shows how well informed policymakers across different countries about five particular issues relating to women and girls.
I want to thank everyone who participated so enthusiastically for joining us this week. These collaborations with the non-profit sector around social impact topics mean a lot to me and, while they take a bit of extra effort to organise and coordinate, they are very rewarding.
Equal Measures 2030 have already sent me a message to let me know how much they enjoyed seeing all the visualisations this week and that they greatly appreciate everyone’s effort.
Collaborating like this does not just help non-profits to crowd-source data visualisation ideas and great ways to communicate their message in different, creative ways, it also helps to bring more awareness to a topic because your data stories are spread across your networks via social media, so THANK YOU ALL!
To help everyone continue their learning around analysis and dataviz, here are two lessons, focusing on general analysis skills and more specific design considerations.
LESSON 1: YOU DON’T NEED TO USE ALL THE DATA
I know it’s tempting to bring every data point into your story, but you don’t actually have to do that, because finding a single focus, for example using one indicator from this week’s dataset or reducing your analysis to a single country across all indicators, can help make your analysis more meaningful.
If you’re struggling to tell a succinct and effective data story, take a step back and ask yourself if you are trying to do too many things all at once. Sure, you may want to cover all aspects of the data, but don’t feel like you have to do so.
It can actually be easier to take a narrower focus at first and once you have found how things ‘hang together’ you can branch out and investigate whether the rest of the dataset behaves the same way or very differently.
I noticed this week that some people chose to use every single data point for each country, which brings the risk of cluttered looking dashboards that can be more difficult to understand for the audience compared to a visualization that focuses on a subset of the data.
So don’t fear a bit of minimalism and reducing the complexity, even if it means that your dashboard does not answer every single possible question. If it instead answers one or a few questions really well and clearly, you have a winner!
LESSON 2: DESIGNING FOR EFFECTIVE COMMUNICATION
When we create data stories, visualisations and dashboards, there is always an audience. Sometimes that audience is yourself, when we just explore a dataset and put some ideas together for our own benefit and other times we have no idea who the audience is. Designing for a known and an unknown audience is quite different.
Designing for an unknown audience
Usually an audience will look at our work and want to gain information and insights from it. With Makeover Monday we publish our work online and we don’t know who the people are who eventually see our visualizations. That means that during the process of creating these dashboards and data stories, we need to keep in mind the end ‘users’.
As we don’t know our entire audience, we should assume the following:
- The user has not spent any time exploring this topic beyond hearing about it in the media, for example. This means we need to assume they have no background knowledge.
- Whoever looks at our visualization is not familiar with interactive data visualizations.
- Our audience has not been given any explanation of acronyms and jargon.
How to help your audience understand your work
What this means for us is that we need to design our visualizations in a way that compensates for the above. How? By…
- Providing sufficient background information on the topic – This doesn’t have to be an essay explaining everything in detail, but it is helpful (and important) to state very clearly what your visualization is about, what your key insights are, and why this matters. The easiest way to provide that information is with an effective title that relates directly to your viz, with a subtitle or short description outlining the ‘problem’ or question you analysed, and some annotations or call-outs that show the insights you found.
- Giving clear instructions on how to interact with your dashboard, especially if interactivity is essential for your audience to find out additional information (tooltips) or parts of the story (filter actions). Including a small element (text, icon, symbol, etc) that encourages interactivity (e.g., ‘click here to filter’ or ‘hover to find out more’) can make a big difference to those people in your audience who are not familiar with interactive dashboards.
- Adding explanations for any technical terms or acronyms you use, as well as the metrics (unless they are genuinely self-explanatory). You can provide those definitions in the text or description, or alternatively add definitions inside the axis labels and annotations.
Be creative and find ways to give your audience the information they need to understand your viz and conclusions. The above are just a few examples to get you started and there are many ways to can avoid confusion.
Is your viz easy to understand?
Test yourself while building your viz, read your title out loud (seriously). Read what the subtitle and descriptions say. Read them aloud and read them slowly. Then ask yourself whether your title and text will make sense to the audience, are easy to understand (using simple, plain language – we’re not doing poetry here) and reflect the essence of your analysis and findings. When you read things aloud you notice much more easily whether something ‘sounds right’. If not, make some changes.
Check whether you have given clear instructions for your users in case you want them to interact with the data.
Lastly, did you define any specific concepts, metrics and assumptions? Will someone with absolutely no background knowledge on the topic be able to understand and follow your analysis, insights and conclusions? Will they feel informed and like they learned something or gained a better, deeper understanding of a specific issue?
Practice makes perfect
It may seem like a big task to get all of these things right every time and it takes time and consistent practice to do so. This week we worked with survey data, which can be very difficult to understand as well as communicate to others. Survey data contains answers to questions that somehow need to be conveyed in our visualization without replicating the entire survey. Extracting the essence of the answers provided by the respondents and visualizing this effectively is difficult to do.
Take every week as an opportunity to implement some of our suggestions. Some datasets lend themselves to interactivity and you can aim to give clear instructions and make interactivity the focus of your viz. Other datasets contain metrics that are more complex, for example when we looked at solar eclipse data last year and had more scientific content which people explained in their dashboards.
Don’t feel like you need to do everything every single time, just keep using these suggestions to tweak things and to identify the areas you’d like to improve on.
A great example this week showing clear instructions as well as comprehensive descriptions of the issue at hand comes from Naledi, who included a very effective legend in the top right corner of her dashboard to explain ‘how to read this chart’.