Feedback needed: Helping users make successful data visualizations

Choosing the right visualization or chart type is important. Research has shown that a majority of dashboards suffer from “Visualization Technique Problem”. This is often a problem of using the wrong chart type for the data. This makes a dashboard and its information difficult to understand. A dashboard that does not communicate its data is not useful.

At DHIS2 we want to help users to create understandable data visualizations. We have been discussing ways we would help users select appropriate chart types for their data. We would like to ask you, the DHIS2 community, for feedback on our ideas.


Display recommendations for visualization/chart type in Data Visualizer app. Recommendations can be both generic and based on the user’s selected data.

Generic recommendations

Display help text alongside chart types. This text should give a short, understandable summary of suitable uses cases.

Displaying the help text as a tooltip (design mockup):

Displaying the help text in the chart selection menu (design mockup):

Possible downsides:

  • All use cases for a chart type will be difficult to communicate in short help text. Users may worry that anything other than what is listed in the help text is invalid. This is not the case.

Recommendations based on data

Display help text/messages when DHIS2 detects that the selected chart type may not be suited to the chart data. For example, if a pie chart is chosen with the time dimension as series. DHIS2 can inform the user that pie charts are not ideal for comparing data over time and can instead recommend a column or line chart. These messages can be dismissed without taking action. They could also be disabled in user settings.

Displaying a recommendation message based on selected data and chart type (design mockup):

Possible downsides:

  • The system must always give correct recommendations. If an incorrect recommendation is given the user will lose trust in all future messages.

  • May be difficult to implement for complex data selections.


We value your feedback. What do you think of these potential features?

  • Do you think they would be helpful to you, or others you have worked with?

  • What might be more helpful, generic recommendations for choosing a chart or specific charts based on your data?

  • Are there any resources you use today to help select the correct chart for your data?

  • What other common data visualization mistakes do you see?

  • Any other feedback or suggestions for how we can help users create valid, understandable data visualizations?

Thank you for your time!

Joe Cooper
DHIS2 User Experience Designer


Hi Joe,

Thank you for following up the research findings, these are nice and thoughtful features :slight_smile: Quickly answering your fourth bullet point, I have detailed data on common visualisation mistakes related to the research. Just let me know if you would like to take a look at them.

Responding to the downsides you listed above:

  • The variations of terms used in data visualisation can complicate how we present the recommendations briefly and concisely (for both features). For example: visualisations - charts - graphs; (visualisation) type - method - technique; variable - data element - indicators (probably known issue: some users struggle to differ indicators from data elements, as they usually use term ‘indicators’ in health sector, for both indicators and data elements)
  • Besides the risk of giving incorrect recommendations, I think creating data visualisations is an art (into some extent). Some indicators can be presented in different ways depending on the purpose of the visualisation (especially the complex ones - you have mentioned this). An alternative is we may need to propose multiple recommendations (or choices) in the message.

Guideline for data visualisation
Do we have a comprehensive and general guideline regarding data visualisation choices and considerations within the platform? This can be added as a link in the recommendation message. So they can consult to the guideline in case they are unsure about following the recommendations.

Still about the guideline, sharing our experience in Indonesia, you probably know this chart below (sorry for the Indonesian version).


We introduced it to our users and most of them did not find it useful as they did not understand how to apply it to their context. Some of them mentioned that it is “too technical” (I would interpret it as “because the guide uses the terms we don’t understand”).

We got these responses for both English and Indonesian versions. This is also an example where terms in data visualisation vary across disciplines/field/sectors. In addition to that, some users suggested if this visual guide can be narrated instead.

Hope this is useful. I will be around OJD 6th floor until Annual Conference, feel free to talk :slight_smile: