The Data Visualization Creation Process

Creating a data visualization is a bit like developing a survey. There’s more to it than just making a list of questions. You’ve got to start with purpose and audience. Otherwise, you might very well end up with something that doesn’t actually meet the need.

For a data visualization to be effective at communicating your key message, you need good data —data that answers your evaluation questions. So the data visualization creation process actually relies upon decisions made along the way of your evaluation journey…. 

Start with end-use strategizing

It all begins with a well-designed evaluation plan where the purpose of each activity is clear in supporting your objective. Where you have figured out what you want to know and how you plan to use the results. Where your data collection instrument is designed to answer the evaluation questions you want to know about. And where the analysis can compare and summarize responses in categories that inform the evaluation questions. 

It’s hard to tell a clear and powerful data story without purpose, clarity, and intent from the beginning. 

Prepare and analyze the data

Once data collection is complete, save a copy of your raw data and start cleaning the copied data. Apply a heat map (with conditional formatting) to the dataset to highlight cells with duplicate values, missing data (empty cells), and outliers (mistakes or unlikely responses outside of the normal range). Try to correct errors if possible. Set a threshold for incomplete data and remove empty cells or incomplete rows.

Review the data several times until you become familiar enough to start seeing patterns. Calculate frequencies, percentages, ranges (highest and lowest values), averages, medians, modes, etc. Use cross tabs to look for differences across categories. Run statistical tests if the data or your audience requires it. Write down your first impressions as well as any questions that arise.

Identify possible data stories

What stands out in the data as important, curious, or surprising? How does the data support or inform your purpose? 

• Does it illustrate the scope of the problem (show magnitude or comparison)?
• Does it expose a lack of awareness or misunderstanding about the issue?
• Does it confirm or confound expectations (e.g., did the problem decrease after implementing a tobacco prevention policy?)
• Does it point to necessary next steps (e.g., the need for more community education)?
• Does it motivate people to action? 

Sometimes the data won’t show a noteworthy pattern, and that’s okay. Not all data needs a visualization.

Select a data story based on purpose and audience 

Choose a chart type that aligns with the function you need — whether that’s comparing categories, showing change over time, illustrating proportions, or highlighting relationships. Experiment by sketching several formats out on paper, trying different layouts, headlines, and labels to see which version best conveys the data story.

Adapt a template

Then search the Data Viz Idea Lab for a template or example that you can adapt for your needs. Apply data visualization design principles when using color, bolding, and primacy to emphasize your data story. (see Data Viz Basics at https://datavizidealab.ucdavis.edu/getting-started)

Test and disseminate your message

Pilot test the data visualization on audiences not familiar with the data to see if your key message is being communicated clearly. Revise as necessary.

Share the data visualization in appropriate forms (e.g., presentations, fact sheets, reports, social media posts, etc.). Provide supplemental context and interpretation as needed, based on audience and form/medium (Does space allow for it? What expectations are there about the level of detail?)

Creating an effective data visualization is more than just choosing a graph option in Excel. It requires thoughtful, deliberate choices, but the payoff is a visualization that communicates your message with power and precision.