Vizuators
Получить консультацию

How to visualize Survey Data in Tableau

You got some survey data. It could be Net Promoter Score survey, marketing research jf potential customers, or employee satisfaction provided by your HR department.
Your task as a data analyst is to provide actionable insights for your team. Not only get it, but present results to decision makers. And Tableau is a great help here. Several common rules and pittfalls are valid for any field. Let's go through them.

Steps of survey data analysis we cover below:

What to consider in survey analysis?

You've done a survey and collected data. What to do with the incomprehensible table?
Data collection is only the middle of the road, and further is the task of correctly interpreting them and conveying insights to decision-makers.

Tableau has rich functionality that will help us bring table data into a form suitable for visualization. We will talk about the possibilities of Tableau and what pitfalls there may be.

With a bit of work to get the data set up correctly and with the right tools to interrogate and visualize your data, you'll get a better, more profound understanding of your decision-makers.

Likert scale, Percent Top Bottom
Steve Wexler, Data Revealations

Survey data
and question types

When we collect data, we design them. It's important to plan carefully what you want to see in the final report before you start the survey. You'll be unable to segment the answers by the audience working field if you didn't ask about it in the survey.

Quantitative research or qualitative research?

What's the difference between them?
Remember
Everything we want to know and the opinion of which part of the audience depends entirely on the survey.

Open-ended and close-ended questions

What kind of data we will collect depends on the questions. Further data cleaning and visualization depend on the questions too.
Question types define features of survey data and dashboard functionality.
Close-ended question
  • Quantitative questions gather numerical data, which can be analyzed using a data analysis platform. It’s a close-ended question generates a limited set of responses.
  • Plus: easy to analyze. High popular Likert scale question or Net Promoter Score is a close-ended question with multiple response questions.
  • Minus: high probability of thoughtlessness of answers, possible random choice, the automatism of the respondent.
Open-ended question
  • Qualitative questions are often open-ended and help answer "why” and gain context about quantifiable data and understand hard-to-quantify behaviors. Open-ended questions require free-form answers. They are justified at the stage of trials, pilotage, determination of the study area.
  • Plus: it allows you to probe deep into the respondent's answers, gaining valuable information about the subject.
  • Minus: respondents need more time to respond. But the main problem for you is further work with the analysis of open-ended questions.

Tableau data cleaning
and preparing

Unless you are fortunate, the data you get from your survey provider will not play nicely with Tableau. We need to clean and reshape the data.

Survey data software and tools


Here it's the popular tools and technologies available to help you get your data set up for visualizing survey data:


  • Tableau's Built-In Pivot Feature
  • Tableau Prep
  • Tableau's Add-In for Microsoft Excel
  • Alteryx Designer, and so on.
Tableau web data connector

Tableau Prep, Alteryx, Tableau Desktop joins

What does their setup look like?

Tableau's pivot feature

Start working with the built-in functionality of Tableau.
An Excel or CSV file from most survey tools will contain one row per respondent and hundreds of columns. Each column corresponds to a different question in the survey.

First, we need to pivot/reshape the data so all questions merge into two columns.

Then we make label responses, numeric responses, logical groupings, and mapping of question IDs, all in a human-readable form.
using pivot tables to analyze survey data

The columns become grouped

Label responses, numeric responses, logical groupings, and mapping of question IDs. By Steve Wexler
Tips
Create a table with helpful metadata for each question ID and group related questions together.

Cartesian product problem

Even though close-ended questions seem to be created for analysis,
there is a problem with processing questions with multiple responses. It's because of using Cartesian join. What's the problem?
  • If we use close-ended questions with single answers (gender, age, place of residence), processing in Tableau is not tricky.
  • But if we work with close-ended and multiple response questions, we bring such answers to a flat table. Cartesian join (product) works for this. A Cartesian join can be helpful when dealing with sparse data or data blends or unpivoting to create a very tall data source to get the desired layout.
  • But in the case of Cartesian join for multiple responses, unwanted data multiplication occurs.
  • Handling multiple answers has become more accessible since the Tableau update 2020.4. Using relationships functionality, handling multiple responses has become much more manageable.
Cartesian join multiple responses
Cartesian join for multiple responses
Unwanted data multiplication

Our Tableau Prep experience: data model and relationships

We worked with two different-scale survey projects. For both cases, we have built a data model. How did she look?

2000 respondents project

We select single answers as a separate stream. Multiple responses were also separated into different data streams and divided into columns, Pivot, and combined.

100 000 respondents survey project

We decided to reshape the data model into a logical data model.

So, single responses became a separate dataset (called Extract 1). From each multiple answers, we also made data sources (Extracts). Further, using logical links, we connected our various responses to single ones – by respondent ID.

Data collection survey tools

Tableau can create insightful visualizations from several data sources without amassing everything into a centralized data warehouse.

Survey data storytelling

Once you've collected all the most pertinent data points using survey analysis tools, it's time to start telling your story.
The data can reveal much with advanced goals and the right questions in your survey. And Tableau has built-in capabilities to look at data in all ways. Users could be excited about looking at research data. Isn't it great?
So, what questions should we start storytelling with?
Demographic questions
Demographic questions are potent tools for segmenting your audience by background and occupation, allowing you to dig further into the data. And when you put it together in a dashboard, survey data allows users to explore and filter survey items based on various demographic profiles.

Questions such as Gender, Location, Generation, Weight, etc., allow you to understand your target audience better. You'll find out the opinion of the particular audience segment you are interested in.
Tips
Use the Count Distinct function COUNTD ([Respondent ID]), a unique field representing each respondent in the data, to return the number of respondents by demographic profile.

Checklist for your survey data to make a story

Tableau gives an ability to answer customer questions on the fly. What to add to the visualization of survey data for best storytelling?
  • Filters
    Filters are the most straightforward tools for finding trends under the surface of your survey data. Filters allow you to see how a particular group of respondents answered the questions in your survey based on how they responded to other questions in your survey.
  • Custom crosstab report
    Crosstabs reports are tables that show how different groups of respondents in your survey (the columns of the table) answer various questions in your survey (the rows of the table). Like filters, they help look under the surface of your top-line results and see how different people answer questions in your survey differently.
  • Metadata
    In survey research, metadata is vital. It describes statistical data from survey conceptualization to data dissemination. Metadata can be comprehensive and encompasses populations, methods, survey instruments, analysis approaches, results of instruments, and so on.
  • Tooltips
    Tooltips give survey users more information about a specific word or phrase in your survey. For example, you can use a tooltip to define a word in your survey people may not be familiar with or give extra context about how to answer a survey question.
Instead of one static infographic, present survey data in an exploratory format. Interactivity will show the relationships among the demographic profiles. But make sure you provide your viewers with the necessary tools and instructions to navigate the data correctly:

  • a parameter control;
  • color highlighting;
  • visual ranking within the category.
highlight table dynamic segments
Even an ordinary table can speak by making it a highlight table
Steve Wexler added to this table also dynamic segments

Visualizing survey data

To communicate your findings in a way that users can understand visually, you need to choose suitable types of charts and graphs. As usual, the choice of visualization is dictated by your data. Use only those types of visualization that clearly show the accents and enhance them yourself.

Here are the most popular and suitable types of survey data visualizations. Help in building.
Tips
Inform the viewer of the margin of error, which is in every survey. It builds trust with your users and convinces them of transparency.

Tableau survey data dashboard

Now it's time to share your story and put visualizations on a dashboard. This format is a very efficient and effective way to communicate survey data – to share insights at the presentations at the meetings.

How to customize your Tableau dashboard based on user requirements?

Filter answer data
Filter data by month or week, and then show charts side-by-side to easily compare scores over time.
Select which questions to show
It's all about saving time for dashboard users. Focus them on key metrics and metric insights.
Customize your layout

Select which worksheets and objects do you want on the dashboard and where to place them. This is where a sketch on paper comes in handy.

Add comments
Add comments below or next to individual charts to flag changes or provide more context.
Follow the corporate style quide
Tailor the look and feel of your dashboards with custom colors and fonts.
Transform your survey results into a report your stakeholders will understand.
Your dashboards dynamically stay up-to-date and can be shared with anyone via a weblink.

Customer satisfaction comparison dashboard example

Pay attention to well-placed charts, the ability to filter and also comments and summary by the author.

Using survey results

Who is the dashboard user, and what is his technical level? Maybe he is not deeply immersed in analysis but are just business users? Maybe he needs not a dashboard,

but a simplified format?

Power users

Allow them:

-to dig into their data using filters in packaged workbooks and the free Tableau Reader application;

-to export chart images for presentations or download crosstabs for their needs.

Business users

Send them:

-a PDF of the workbook;

-specific images of particular slides for their presentations.

Tips
We recommend simplifying your planned dashboards if you have a lot of data in databases. Remove unnecessary filtering, "lighten it", and raise productivity.

Analyse survey data with us

One-hour free consultation with a Tableau business analyst and developer. Determine the main objectives of the project, readiness of data sources, and ask any questions about Tableau.

By clicking the button, you agree to the processing of personal data. We use it to stay connected. Read more