Data visualization and storytelling


Five ways to effectively visualize survey responses

On a scale of one to ten, how do you feel about visualizing survey data?

Data exists in myriad forms, and a common subject of visualizations in today’s information-rich world is survey results. These can be complex to illustrate clearly, however — this post will walk through five ways to display survey data with Flourish templates.

1. Our topic’s namesake — the Survey template

This template is ideal when your survey data contains one row per respondent (best for datasets with 10,000 rows or fewer — larger datasets will take longer to load). Its features allow users to compare respondents’ characteristics and question responses by group, shade, or geographic location (for categorical variables) and size (for continuous variables). You can visualize your data in dots, squares and custom shapes and create everything from scaled circles to a waffle chart. Below is an example using data on some of the first members of the UK’s “Windrush generation”. Learn more by going through the slides below:

Create your own survey visualization now »

Other survey data may have a more aggregated structure that works better with other templates. Teams at Imperial College London and YouGov recently partnered to create a series of charts to display responses to their joint COVID-19 behaviours questionnaire, which we used as inspiration for the next two visualization types.

2. Show Likert scales by group with stacked bar charts

Likert scales, named after their creator, psychologist Rensis Likert, are used in many types of survey research. A common way to present these scales visually is to use stacked proportional bar or column charts to illustrate respondents’ extent of agreement with a given statement. For the Imperial/YouGov project, the team made stacked column charts with our “Line, bar, pie” template to indicate the proportion of survey respondents who disagreed or agreed with certain statements related to COVID-19. Here’s an example:

Comparing the levels of blue and red shading in each country surveyed, as well as the means in light blue, can give an overview of where people are more concerned about catching the virus. Here, it looks like survey respondents from Asian countries tended to perceive COVID-19 as more dangerous.

Create your own Likert chart now »

3. Show multiple choice responses with heatmaps

In some cases, a survey multiple choice question allows respondents to tick multiple options. The Imperial/YouGov team chose our Heatmap template to illustrate the proportions of those respondents who ticked each option across countries and age groups for a question about people who had not been vaccinated against COVID-19. Using the filter control, users can discover statistics such as which groups were more likely to be concerned about vaccine side effects or their effectiveness. Check out our comprehensive guide to the Heatmap template to create your own.

Create your own heatmap now »

4. Show quantitative responses across regions with maps

You can also display data from individual survey questions, especially those with quantitative responses, on a choropleth map. Using sequential or diverging color scales, you can illustrate how people from different countries or regions differ in opinions or perspectives. Learn how to add custom colors and scales to your map here. The example below shows how life satisfaction (measured on a scale from 1-10) is distributed across countries. Countries with no data are shaded in a grey and white pattern.

Create your own choropleth map now »

5. Display proportions of categorical responses with pictograms

Our recently launched Pictogram template allows you to illustrate respondent categories and proportions by designating an icon to represent a numerical value. For example, here’s a visualization showing the numbers and proportions of businesses by industry and total number of employees:

Create your own pictogram now »

Other chart types

There are many other ways to display survey data that we haven’t discussed here. For example, making pie or donut charts using our “Line, bar, pie” template can allow you to display a handful of categories by proportion of survey respondents. And if your data are qualitative? You can use our new “Word cloud” template to assemble what words respondents use most often when describing a certain topic. Here are some examples:

Create your own visualization now »

Over to you!

If you have survey visualizations you’ve created, reach out to to share your creations!