AMA with The Economist's data team

Conversations with Data: #32

Do you want to receive Conversations with Data? Subscribe

Conversations W Data header

In 1843, The Economist’s inaugural edition went to print with a table front and centre. Clearly ahead of his time, the editor of the day recognised the power of data journalism over a 100 years before the field’s modern emergence.

Almost 176 years later and the outlet’s appetite for data driven stories is still going strong. In 2015, they brought in a specialised data team and, this year, they launched a dedicated data section in print.

To find out more about The Economist’s affinity with data, we let you pose your burning questions to the data team themselves. Here’s what they had to say!

What you asked

Let’s start with your data visualisations. How do you decide what to use in your publication?

Matt McLean, Visual Data Journalist: “Mainly by looking at the data. The data structure and type will suggest certain visualisation devices (bars for stock, lines for flow; scatter to show a relationship between two variables etc), but it is always worth experimenting to see if you can find new and engaging ways to visualise the data that go beyond standard chart types but still preserve the integrity of the numbers. We also have to consider what will work in the space available and make sure there is enough variation to keep things interesting -- no-one wants wall-to-wall bar charts.”

Dailycharts

The team uses varied types of visualisations in their daily charts to engage readers with the data at hand.

You’ve being using GitHub to share data from your Graphic Detail section, and it is much appreciated! Will you eventually be sharing the data and/or code from your daily charts as well?

Evan Hensleigh, Visual Data Journalist: “We’re glad people are using the data we’ve published on GitHub -- it’s been really great to see our data reused in new and interesting ways. Unlike our Graphic Detail pages, our daily charts usually use data that is already generally available, and don’t involve a lot of original analysis, so it doesn’t make sense for us to re-publish the data. But in those cases where our daily charts do involve us creating an interesting dataset, we may start publishing those datasets for other people to do more work on.”

And what other open data steps will you be taking this year?

“We’re looking for ways to make our data easier to access and use. We’re currently working on creating an R package with our publicly available data that will make it easier for people to access our data, mimic our styles, and use things like the Big Mac Index in their own projects.”

What is your dream dataset (that you don't already possess)?

Marie Segger, Data Journalist: “Generally we try to be imaginative and if we don’t have the data we need, we try to get it by either asking for it or by scraping the web as one of our team members has described in an article on Medium. Personally, I think that the relatively new Missing Numbers project that highlights data that the UK government should be collecting but isn’t, is a great place to look for interesting ideas.”

James Fransham, Data Correspondent: “I think Raj Chetty, the Princeton economist, has done incredible things with microdata, constructing longitudinal studies using data and methods that no other social scientist thought possible. His website has a wealth of information, and I would love to have the keys to public data that would allow us to mimic his research outside of America.”

G. Elliott Morris, Data Journalist: “I would love to get my hands on Barack Obama’s political data operation, as unveiled in Sasha Isenberg’s The Victory Lab. As described, the data contained voting records for hundreds of millions of Americans and was compiled by on-the-ground staffers who went door-to-door asking them about anything from their political preferences to dreams for society. It is the biggest of political ‘big data’, and we could use it to improve our already-stellar modelling interactive.”

The Economist Visual

A screenshot from the team’s interactive statistical model, which predicts voting preferences based on different factors.

On the topic of modelling, how do you build predictive models for sustainable development using the SDGs framework?

James Fransham: “A great question. I was in Ghana recently with the UN Foundation exploring how the government there is measuring poverty for the SDGs. The scale and level of ambition of the government is impressive: attempting a fully-digitised census next year; produced a digital postcode for every 5mx5m piece of land; and a biometric ID for each citizen. Nonetheless, data gathering remains a time consuming, expensive and exhausting process. How we build a model that could anticipate Ghana’s progress is meeting the SDGs? I think first we grab every bit of historic data we have about the country, and other similar countries at different levels of development, throw it into a big soup, and hope that something tasty comes out!”

You can keep up-to-date with The Economist’s data team and their latest work here.

Our next conversation

From an established outlet, to something more grassroots, for our next conversation we’ll be heading down to Southeast Asia. The region is booming with data journalism communities, emerging through meetups and local learning groups. To hear all about it, we’ll be hosting an AMA with Kuek Ser Kuang Keng, founder of DataN and co-organiser of Hacks/Hackers Kuala Lumpur. Comment with your questions, or to let us know what you'd like to see in future editions!

Until next time,

Madolyn from the EJC Data team

subscribe figure