NZ Research Trends
Data visualisation that reveals the NZ research workforce’s movements
Duration
6 weeks
Team
Ashleigh Kinnear Eva Zhou Kayla Palmer
My role
Data analysis, ideation, UI design, interaction design
Toolkit
Excel, Figma
Project background
Our client Te Ara Paerangi Future Pathways is a multi-year programme under the New Zealand Ministry of Business, Innovation & Employment (MBIE), which focuses on the future of New Zealand’s research, science, and innovation system. Through Te Ara Paerangi Future Pathways, we are building a modern, future-focused research system for New Zealand to meet the challenges and make the most of the opportunities ahead of us. Part of this programme includes a review of how research is currently being organised and funded. New Zealand’s research system was designed nearly 30 years ago, and some parts of the system are not working as well as they should be.
Overview
Objective:
- To analyse New Zealand’s research workforce based on a data set provided by Dragonfly
- To search for findings and insights through the analysis
- To visualise the data set to help inform the development of a new research, science, and innovation system
Analysis and scope
- Find out trends and patterns of the research field in NZ and overseas and visualise it
Initial data analysis
We did some early explorations in Excel to look at the data set from different perspectives. We tried multiple approaches, such as breaking down the data by different time frames, subjects, quantity of papers and countries. Through this, we gained an overall and comprehensive understanding of the data. We also compared the new diagrams we had with the existing diagrams from the original Dragonfly report, to seek new findings and insights that can be built upon the existing research.
Finding trends and insights
We started by finding some really simple numbers, and found it interesting to see how many people have moved to New Zealand to publish, and in comparison how many New Zealanders had left. We then decided to dig deeper into the idea of ‘movement’ and tried to find some more interesting data. We narrowed in on a few areas of focus, outlined below. We decided to measure these by the total number of papers published rather than individual publishers, to account for some of the bugs in the dataset.
Insight 1
- Due to the large size of the range of the data, the trends found in Excel were visualised in a way that was challenging to read. This is mainly because the gap between the highest and the lowest country is too large. For example, the highest amount was 10,104 papers by New Zealanders returning to publish after leaving,
and the lowest was 1 paper published in New Zealand by a publisher from Costa Rica and a few other countries. As a result, the majority of the countries with smaller numbers almost disappear from the diagram.
Insight 2
- The data default displayed in alphabetical order, which made it hard to make comparisons between the countries unless you knew the exact values of each bar. This caused a large cognitive load even for us who were getting used to the data at this point.
Challenges
How might we provide an efficient way to make quick comparisons of the quantities of published papers among the countries?
The challenges we faced in presenting the data:
- Authenticity is important – include all countries that are collected in the data set
- Readability – How to showcase the small numbers? How to visually balance the high numbers and the low numbers?
- Efficiency – How to expand all the country codes? How to display the paper quantity intuitively?
As the picture shown on the left, our initial attempts in Excel has a few problems:
- The countries with low quantities of paper are hard to identify
- The country codes are hard to read and understand
- The diagram is visually unbalanced
Visual Solutions
Our first idea to solve this was to display the data in chunks so that none of the bars got too small to read. This meant 4 different graphs all displaying what countries people had come from to study in NZ. We started coming up with a way that the viewer could choose what part of the giant graph they wanted to see, and still understand that the 4 separate graphs were part of the same thing.
Layout design
Strategy to present the data:
- Four groups of data in different quantity ranges
- In each group, assign a hover effect to each country to expand the country name and quantity of papers
- Assign an indicating bar to each group, to help put the diagrams in perspective of the scale of the quantity
High-fi prototype
This is what all 4 finished graphs looked like for usability tests. The four graphs are combined as one diagram in Smart Animation so that users can switch in between each other.
Usability Test
Finding 1:
- Overall, users perceived the functions of the data visualisation well.
Finding 2:
- Users found the hovering function intuitive to use.
Finding 3:
- Some participants found it a bit hard to search for a particular country using the country code.
Finding 4:
- Due to the space limitations on the screen, some of the country codes and subject abbreviations needed to be displayed vertically and this created inconvenience for reading.
Finding 5:
- Some participants were confused by the indicating bar under the data column bars, expecting the indicating to be interactive because it has the same visual style as the column bars.
Prototype refining
In response to the findings from our usability tests, we made some refinements:
- Proximity – Moved the indicating bar under the buttons, so as to reinforce the interaction between the two when the users switch between the buttons.
- The hovering effect makes a remedy to the vertical country codes and helps clarify the information when the audience points at a certain one.
Final design display
Reflection and next step
We acknowledge the usefulness of sorting the data in country codes alphabetically.
We also acknowledge the efficiency of sorting the data by quantities of paper.
Our next step would be to build a data system with two combined search functions: by country codes in alphabet order and by paper quantity ranking. This will provide more flexible options for users to choose their preferred search method according to their needs.
Some of the limitations of the project include:
- A substantial amount of missing data entries of countries from the data set collected by the current AI system (about 15%)
- Due to missing data, some personnel movements were omitted and can only be traced by other means
- The missing data may lead to disproportions of paper counts in NZ and abroad