A data driven analysis on the results of the Malaysian Elections, exploring factors such as age and race, that could have led to Malaysia’s first ever hung parliament.
As a team, we learnt (and got better) at doing more in-depth, data driven, analysis with a relatively quick turnaround rate compared to the rest of our interactives. This project was one of the more heavily data-driven stories the team has done. We got to experiment a lot more (especially for chart designs) using different types of charts, maps, writing good and clear annotations. This would be helpful for future data-driven analysis and upcoming elections stories.
“Due to the relatively quick turnaround rate of this project given the nature of the topic, we used simple tools like datawrapper and svg for most of the charts. For the map scrolly, we had a template that we used for [another project](https://www.straitstimes.com/multimedia/graphics/2022/10/malaysia-ge2022-government/index.html) we did prior to the Malaysia elections, so we repurposed that component and updated the data.
For the scrolly, we chose maps because it provides spatial context and would allow readers to visualise the areas which were won by each party. The bar graph at the top of the map was also added to provided more context. Each square represents one parliamentary seat, and is colored according to the winning party for that seat. Maps alone can be quite tricky because a large area doesn’t necessarily mean more seats in Parliament. For example, Sabah and Sarawak combined are larger than Peninsular Malaysia in area but have less seats in parliament. So the bar graph helps to add more context.”
Context about the project:
The hardest part of the project was ensuring that we were carefully wording what our data was repesenting. Age was one of the factors that contributed to the divided polls in Malaysia’s latest election. While we had data broken down by age for each constituency, this data was based on all registered voters. In Malaysia, everyone who is eligible to vote will be considered part of the electoral roll under the automatic voter registration system. But in reality, not all registered voters actually turned up to vote. We did not have data for the turnout rate, as this would usually be only made available a couple of months later, and they are not usually available to the public. Hence, when doing our analysis, we had to make it clear in our charts and text that our analysis was based on registered voters’ profiles at the constituency level. We also spoke to several analysts to ensure that we were accurately presenting our findings.
What can other journalists learn from this project?
I think journalists can learn that even though we might not have the ideal dataset (in this case, the voter profiles of those who turned up to vote) we would have wanted, it does not mean that our dataset would be entirely useless. We were still able to get meaningful insights, trends and indications with the registered profiles dataset.
We also started this project wanting to focus on age alone, but then down the road we realised that the dividing factors weren’t clear cut and that there were several reasons, beyond age, as to why voters chose to vote for Perikatan Nasional (PN). Another factor was race, which is more contentious but important, and would not have been entirely balanced to leave it out in our story. It was helpful to speak to experts on the matter. For example, for the race factor, we spoke to Dr Welsh who has been doing research in this field for several years. With her experitse and knowlege, we were able to talk about how while PN appeals to many young voters, they are mainly young Malay voters. And of course, we also had to back her findings up with quotes from other experts and analysts.