Webinar on composite indicators and data visualisation

On Wednesday 29th June we were very pleased to launch our new composite indicator group and website with a webinar. This was a great success, with over 100 people registered from a variety of organisations. To everyone who registered, watched and/or helped spread the word about this, a massive thank you!

If you missed the event or would like a recap, you can find a recording online here. Slides are also available here.

The webinar was intended to give a short overview of composite indicators and data visualisation, followed by a quick tour of a couple of projects that we have worked on. We opened by introducing who we are as a group, who we are as individuals, and we gave overview of the services that we offer.

Why composite indicators?

In the next presentation, we explained why composite indicators are used and their main advantages. The key points were:

  • Composite indicators (CIs) tend to tackle complex multidimensional issues that can’t be directly measured, and are possibly difficult to define clearly (see our recent blog post on this subject)
  • A multidimensional concept is represented by a CI by deconstructing the concept into dimensions and sub-dimensions, followed by reconstructing these chunks using indicators.
  • The main advantages of CIs are:
    • They can make a complex issue accessible to non-experts
    • They act as an entry point to explore the underlying data set
    • They can identify high level trends and relationships with other variables and concepts
    • They give clear, simple and strong messages, which get picked up by the media. In this sense they can be a powerful advocacy tool.
  • A composite indicator never substitutes the underlying indicators and data; rather it is complementary to it.
A summary of the steps we follow in building a composite indicator

We then gave an overview of the main phases we follow from the beginning to the end of building a composite indicator. This begins with expert and stakeholder consultation, then moves to data collection and processing, and then finally to visualisation and presentation.

Why data visualisation?

Source: https://www.thecommonwealth.io/covid19dashboard/#/

Here we explored some general concepts in data visualisation. The main takeaways were:

  • Visualising data can identify trends, patterns and differences that would be hidden by descriptive statistics
  • Simple concepts such as colours, when used carefully, are powerful tools for highlighting patterns and structure in the data
  • The main aim of data visualisation is to help the user to make sense the data – often this is best achieved with simple and clean graphics
  • But, if visualisations can also be fun and pretty, that also makes them more interesting!
  • Visualisations can bring out trends and patterns, but can also be used to provoke emotions.

Composite indicators and data visualisation

High-quality data visualisations (usually in the form of a web platform) are a great asset for composite indicators. In our opinion, interactive visualisations:

  • Give a visually striking and accessible home for your data and analysis, which can be continually updated
  • Let users explore the data in the way that they want: different users find answers to different questions
  • Let users drill down into the hierarchical data set
  • Allow detailed comparisons between countries (or other units)
  • Facilitate economic and demographic analysis, and trend analysis
  • Give a platform for advocacy and communication

Case studies

We then presented two case studies.

The Quality Infrastructure for Sustainable Development Index was developed with the United Nations Industrial Development Organisation. It measures quality infrastructure and its interactions with sustainable development, at the national level. We constructed this index from the ground up, following expert and stakeholder consultation, data collection, cleaning and processing, and finally static and interactive visualisation at the web platform that can be found here. Details can also be found in our portfolio.

The Water Energy Food Nexus Index was built to measure the availability and access of water, food and energy, at the national level. We performed the data processing and analysis, and built the web-platform which can be found here. Details can also be found in our portfolio.


Our webinar ended with a Q&A. Some excellent questions were raised, which we summarise very briefly here.

Q: How do you communicate uncertainty and risk in composite indicators?

This is a great question. Composite indicators have many sources of uncertainty, including indicator selection, data, methodological uncertainties, and so on. We can estimate the uncertainty on composite indicator scores and ranks using uncertainty analysis and sensitivity analysis. This requires rebuilding the composite indicator many times, randomly varying assumptions. This is a little complex but can be done automatically using the COINr package in R.

Q: How well do visualisations work with mobile devices?

Our websites are carefully designed to work responsively on mobile/tablets as well as desktop screens. Charts are selected with mobile users in mind – an example being the globe visualisation in the WEF Nexus Index.

Q: Which tools do you use for interactive data visualisations?

We use tailor-made data visualisations built using JavaScript, HTML and CSS. We leverage JavaScript libraries such as d3.js and vue.js. We do not use commercial tools so there are no licencing fees!

Q: Can we help with certain parts of composite indicator development and analysis?

Yes please get in touch with us if you only need help with certain parts of the process, rather than the full package.

Q: How do you aggregate indicators?

The fundamental steps to aggregating indicators are:

  1. Indicators must be cleaned
  2. Indicators are “normalised” which means bringing them onto similar scales
  3. Aggregation is performed following the hierarchical map of the concept, aggregating indicators by groups using (usually) weighted means. In the simplest case, we use the weighted arithmetic mean.

This misses out many other important steps such as multivariate analysis, outlier treatment, indicator and unit screening, and so on. There are also many many ways to weight indicators.

We will publish more blog posts in the future dealing with these issues. Subscribe to our blog to keep updated!


We are very thankful to everyone who attended the webinar. Again, a recording is available here, and slides here.

If you have any thoughts or questions, feel free to leave a comment below. Or, contact us! We’d like to run another webinar later in the year, so stay tuned for that and let us know if there is anything you’d like us to talk about in particular.


Subscribe to our blog

More Posts

Composite indicator map

What is a composite indicator?

Good question. The short answer is that a composite indicator is a mathematical aggregation of a set of indicators, usually aiming to measure a multidimensional