Project description

I am a Visual Projects Editor working in The Guardian’s Visuals team, where I report, design and code in order to break and explain stories. My broad skill set allows me to work on projects from conception through to publication, crafting innovative forms of digital journalism in which text and visuals work seamlessly together.

The stories I have chosen to represent my portfolio cover the full spectrum of data journalism in a newsroom: static, reactive charts for the US midterms, a full live election tracker for the UK local elections, original data analyses that reveal big stories and interactive data visualisations that push the boundaries of data-driven storytelling.

Utilising my software engineering background, I have analysed large, complex datasets such as the Brexit voting records of all 650 MPs. This led to the creation of over 4 million “similarity scores” for pairs of MPs across 10 key Brexit votes, ultimately revealing the emergence of four cross-party factions. I then worked this into an interactive data visualisation that stepped readers through the fracturing of British political parties over the last 9 months. This network analysis correctly categorised the 7 Labour MPs who went on to quit the party and form The Independent Group.

I also revealed that more than 1 in 12 stores have been lost from English high streets in just five years. I analysed multiple years of Ordnance Survey’s “Points of Interest” database, featuring more than 4 million public and private businesses.

This year has been a hugely exciting year for the Guardian Visuals team, and these select projects of mine are the result of many hours of experimentation, exploration and refinement.

What makes this project innovative?

Many of the projects utilise software engineering techniques to analyse huge datasets either in JavaScript using node.js, R or QGIS. For example, for the “High Streets Crisis Deepens” project, I worked with Ordnance Survey’s ‘Points of Interest’ directory, which is a geospatial database that contains over 6 million businesses across the UK. Using a nearest-neighbour analysis in QGIS I identified the densest shopping streets that make up town centre retail, creating a focused analysis of UK retail that specifically excludes shopping centres and out of town retail parks. I also spend a lot of time working on how best to communicate the full results of my analyses to readers in an accessible, memorable manner. Doing this often requires tearing-up the data visualisation rulebook in order to craft bespoke visualisations that are tailored to the exact insights we’re trying to convey to our readers. For example, in the “Trump Teleprompter Test” piece, I integrated short video clips of Donald Trump speaking at his campaign rallies into the data visualisation in order to audibly reinforce the difference in Trump’s language between his on and off teleprompter speeches. For the Guardian’s gender pay gap coverage I worked on the idea of a ‘calendar’ which would mark how many days women are working for free in the 7,853 large British companies where there is a pay gap that favours men. This is another example of how unusual data visualisation techniques can memorably convey key stories within datasets. One technique that I employ regularly is the use of ‘scrollytelling’. This worked incredibly well for the “How Amazon Became The World's Most Valuable Retailer” interactive, where we wanted to demonstrate just how rapid Amazon’s market value has grown. By designing this bespoke visualisation, we turned what is fundamentally a business line chart into a widely read interactive that sparked a widespread discussion about the company’s mammoth size.

What was the impact of your project? How did you measure it?

Our metrics for success vary from project to project. For the “how Brexit revealed four new political factions” interactive we saw a wide discussion on social media, with many people choosing to highlight their own MP’s voting record, often @ing their MP in the tweet. This kind of response is exactly what we were hoping for when we started publishing our MP vote trackers, since greater transparency leads to better informed citizens – crucial in the age of misinformation and spin online. This piece was extremely well-read with hundreds of thousands of page views and above-average attention time for a piece of its length. For the “High Streets Crisis Deepens” project we wanted to expose towns and cities that were struggling the most. My data analysis went on to inform a number of on-the-ground reported features, with the piece from Sheffield attracting the attention of Sheffield City Council and prompting a response defending their development plans for the city.

Source and methodology

I have sourced data from a number of open data platforms, such as government sites and the World Bank, as well as partnering with companies such as Ordnance Survey to use their commercial datasets for data-driven stories. For events such as the US midterms and the UK local elections, I wrote custom node.js scripts that could clean, analyse and visualise live datasets in a matter of seconds. We would then take the resulting SVGs into Adobe Illustrator and add final annotations, massively reducing the time taken to produce a graphic on results day by preparing data analysis and visualisation scripts ahead of time.

Technologies Used

Most of my projects utilise D3 for interactive data visualisation, R and QGIS for data exploration and Adobe Illustrator for the creation of static graphics.

Project members

The entire Guardian Visuals team has contributed to the success of these projects by offering their own opinions, technical expertise and years of experience to help shape my ideas into the final pieces that are published.

Link

Additional links

Followers

Click Follow to keep up with the evolution of this project:
you will receive a notification anytime the project leader updates the project page.