A Flat Delivery of Dutch Commodity Data
My final Udacity assignment has tasked me to critique an existing data visualisation.
What makes a data visualisation impactful?
Objective: suggest how the Netherland’s Centre for the Promotion of Imports (CBI) could improve an existing visualisation; What is the demand for cocoa on the European market?
What is the goal of CBI’s visualisation?
The CBI article is trying to promote the opportunity of the European cocoa trade to private and public investors. To achieve this goal, the article uses empirical data and simple charts to articulate the positive trend in the market
What does it do well?
Bar charts are an effective way to describe numerical data to a Layman audience. The supporting text is exhaustive and laden with headline statistics and take away messages. The callout boxes breakup the text and help to keep the audience reading — nice.
Where does it fall flat?
- White Space. In my opinion, there isn’t enough. This is a detailed topic. Instead, annotated charts could be used to replace the paragraphs used to give explicit statistics. By doing this, the additional, more detailed information could have been positioned at the foot of the article for those who wish to study a particular topic further
“See the appendix for more detail.” Visualisation Author
- Deliver the Trends. The decision to use ‘years’ as a category rather than a series makes it more difficult to interpret any trends over time. As a result, a side by side comparison of countries’ data over time is less intuitive. Instead, a continuous time series along an x axis would deliver clearer trends to the audience.
- Visualisation of the take home messages. There are several interesting facts given in the text that could have been backed-up with a visualisation. “Between period x to y the data says z” should be partnered with a chart. The audience cannot see the message.
- Allow the audience to interact. This is an online article — not a hardcopy. All visualisations are static. It would be more engaging if the charts included editable filters, annotations or story points to walk the audience through the data.
- Compare to other datasets: The article cites facts from second sources of data. For example, Prodcom data concerning EU 27 member states’ annual production of chocolate products. It would have been interesting to join the two datasets to visualise correlation between the two measures (EU imports of cocoa beans vs EU production of chocolate).
Visualisation 2.0
I have produced 3 visualisations:
- A dashboard: global cocoa imports by region over time
Overall, I attempt to communicate the evolution and trends that may be found within the data. To do this I have used time as a series rather than a category. In my opinion, this delivers the message that the volume of the European market has grown over time (a key message) in a manner that is easier for the audience to interpret.
Second, I add interactive elements to keep the audience engaged. It is now possible for the reader to filter the data and hover-over annotations to explore the data.
“Give them buttons, they will click them.” software engineer, previous employer
- A data story: annual snapshots of each year’s data
Next, I create a data story points to allow the user to actively navigate through the chart and observe the takeaway messages from the article’s text. By doing so, the reader is encouraged to interact with the data. This may be more engaging to the reader.
- Further study: using a second data set, I highlight an additional question; which factor(s) explain the fluctuation of a region’s annual import volume
Finally, I have joined a second data source that is discussed in the original article, Prodcom data, concerning EU 27 member states’ annual production of chocolate products. Interestingly, there is a correlation between the two metrics. However, there does not appear to be an immediate impact on fluctuating EU Chocolate production volume to the Regional annual import of cocoa beans. It is likely that the commodity trading business has a Macro view of regional demand. Clearly, there is a more complex relationship of regional import volumes than the local domestic demand.
