Before choosing software
Dashboards can be created using a variety of software (although the availability of software will vary by department).
Software choices should be made in line with the GDS guidance on choosing technology. Before deciding which software to use, think about:
- the skills of your analysts
- the resource available to build to your dashboard
- where you want to host your dashboard
- how you will meet accessibility requirements
Where possible, you should:
- make use of open-source software
- avoid being bound to a specific vendor
- use common government technology components
- share with other departments to reduce overall cost to government
Skills
Software skills of analysts vary within and across departments. You should, therefore, consider if your dashboard can be supported long-term. Dashboards are live products that require ongoing maintenance and may need updating or retiring at any time.
Cost and hosting
Software licencing can be expensive. Each department will have different software availability and governance around procuring appropriate software. You should:
- minimise the cost of ownership
- reduce the chance of being tied into long contracts or licences
Common open-source software (such as R, and Python) will be available in many departments that support Reproducible Analytical Pipelines (RAP). Bespoke dashboarding software (such as Power BI) may have associated costs.
Consider other costs, such as:
- hosting server fees
- maintenance for publishing dashboards both internally and externally
Consult infrastructure teams before choosing software to ensure your plans for hosting and sharing the dashboard are in accordance with your department’s data security policies.
Host server options available to you will depend on your software choice, along with your department’s IT setup. Some departments can publish dashboards externally, whereas others only have capability for internal publication. Contact your department’s IT team for more information.
Accessibility
External dashboards must meet WCAG 2.2 guidelines, and internal dashboards should also aim to meet these standards for internal users. Even if your users do not currently have accessibility needs, this may not always be the case. Test your dashboard to check your software meets accessibility requirements.
Open-source languages (such as R, and Python) will give greater control over accessibility, but require greater understanding of the software to implement.
Out-of-the-box options (such as Power BI) come with some accessibility features built in, but they can be difficult to customise (for example, changing the markup of content can be complicated). As a result, accessibility features are available but are not always used by analysts in a way that meets the WCAG 2.2 requirements. This means that users who are reliant on keyboards or assistive software can find it difficult to interact with these dashboards.
User load
User load refers to the number of users accessing your dashboard at the same time. You should consider user load when choosing your dashboard software. Test the system under heavy usage to ensure the dashboard stays responsive and stable, even during busy periods.
In an R Shiny app, using asynchronous processing (like future or promises for non-blocking operations) can help prevent the app from freezing or slowing down when multiple users are active. You could also consider deploying the app to multiple URLs to ease the traffic flowing into one location.
For out-of-the-box options, you can control app efficiency during ‘point-and-click’ methods. For example, the “Enable Load” function in Power BI will help you determine if a query’s result is loaded as a table into the data model.
R Shiny and Python (for example, Dash)
Benefits
R shiny and Python:
- are fully customisable – they can look and behave how you want
- are supported by packages like shinyGovstyle and the GDS service toolkit to give dashboards a common GOV.UK look and feel
- are likely to be available for analysts to use, as R and Python are popular open-source code languages used across government
- use open-source code, which makes it easy to reuse and share templates and best practice
- have a large community support base
- work well with large datasets
- can automate testing and quality assurance
- have the flexibility to meet accessibility needs due to control over elements
Limitations
R shiny and Python:
- can involve a steep learning curve for moving from basic R and Python to building R Shiny and Python Dash applications
- require analysts that are skilled in coding to build and then maintain, including managing and updating dependencies
- have some additional interactive components that are not always accessible, and some extensions that are accessible (like HighCharts) may require licences at additional cost – see the HighCharts accessibility module for more information
- may require additional knowledge of CSS or JavaScript to make full use of functionality
Useful resources for R shiny and Python
Resources include:
Quarto and R Markdown
Benefits
Quarto and R Markdown:
- are likely to be available for analysts to use, as R and Python are popular open-source coding languages used across government
- use open-source code, which makes it easy to share templates and best practice
- can extend functionality and are customisable to user requirements
- can automate testing and quality assurance
- have the flexibility to meet accessibility needs due to control over elements
- can either be hosted online or shared as a HTML document if hosting is not an option
Limitations
Quarto and R Markdown:
- may not handle large datasets well when embedded into static HTMLs (consider hosting your dashboard somewhere it can connect to APIs if this is the case)
- do not allow access control for dashboards shared as a HTML document
- have limited interactivity – things that are simple to create in other tools like Shiny, Dash or Power BI may need trickier workarounds to recreate in Quarto or R Markdown
Useful resources for Quarto and R markdown
Resources include:
Power BI
Benefits
Power BI:
- has an intuitive point and click interface, which makes it relatively easy for analysts without a coding background to pick up – the interface has a familiar Microsoft Office look and feel to it
- offers a quick way of getting basic dashboards up and running, so it can be good for working on simple products at short notice internally
- can handle large data sets well
Limitations
Power BI:
- is not open source, so it can be harder to share templates and code across teams for reproducibility
- relies on having data structured in the correct way for interactivity to work as expected
- needs understanding of Power Query and DAX to exploit full features
- has limited customisation and chart options
- is limited in the way it can work with Git for version control
- is complicated for automated QA and testing
- has some built-in accessibility options, but limited control over accessibility as dashboard components are “out of the box”
- is tied into Microsoft’s licencing, and product strategy, which means you are bound to a specific vendor
- has a limited Accessibility Checker which does not cover all possible issues, so additional manual testing is recommended
- lacks built-in accessibility support for custom visuals, which means additional adjustments can be required
Useful resources for Power BI
Complete the Microsoft tutorial on getting started with Power BI.
Other software
More complex dashboarding tools may be appropriate for your dashboard. Languages like HTML and JavaScript can be used to create bespoke dashboards. We advise consulting software developers in your department and working with them to develop these if the other options described in this guidance do not meet your needs.
If there are other software options missing from this guidance, please email Analysis.Function@ons.gov.uk to discuss adding it to this page.
Helpful links
If you have any further questions or want to discuss updates to this page, you can:
- email Function@ons.gov.uk
- join our Basecamp project for accessibility for statistics and analysis or presentation and dissemination of statistics and analysis – if you are not on the Cross Government Basecamp, email Function@ons.gov.uk for an invite link
- contact your departmental presentation or web dissemination champion
- join the #design Channel on UK Government Digital Slack
- join the #content Channel on UK Government Digital Slack
- join the #chat-data-vis channel on Gov Data Science Slack