Reuters COVID-19 Global Tracker

Country/area: Singapore

Organisation: Reuters

Organisation size: Big

Publication date: 8 Sep 2020

Credit: Gurman Bhatia, Prasanta Kumar Dutta, Jon McClure, the Reuters COVID Data Team

Project description:

Getting timely information about the spread of the coronavirus and putting the mountain of statistics the pandemic generates in context for readers is at the core of public service journalism. The COVID-19 Global Tracker gives readers a key tool to filter down the firehose using custom algorithms to anticipate the headlines and responsibly mark milestones, progress and resurgence in a global and local context in five languages. The tool transparently explains the statistics and the limits of the data and ultimately gives readers in 213 countries a way to hold their governments to account for their handling of the pandemic.

Impact reached:

The Global Tracker has become a key tool for our readers to understand the pandemic by providing a reference point for each of the 213 countries we create tracking pages for. 


As the project continues, the audience has consistently grown to millions of readers every month, with especially strong growth among Japanese, French and Spanish-speakers.


The tool helps direct our global newsroom’s coverage by highlighting important developments in every country we track, especially those where local media may be less well positioned to keep track of the daily statistics.


As we move into a new phase of the pandemic, the project has a strong, consistent audience watching the rollout of vaccines in their own countries, and new data we’re currently developing will give readers key information about their access to those programs, locally.

Techniques/technologies used:

The Global Tracker is built entirely using serverless cloud technologies to make the project easy to maintain, especially given the scale of the data it analyzes and presents. The pages and underlying data are republished automatically many times each day to keep up with rolling data collection in Reuters newsrooms across the globe. Those pipelines are built with strong safeguards to make sure the data we’re publishing is as accurate as it can be and also helps identify irregular reporting in our original sources for our newsroom.


In terms of our analysis tools, ours was one of the first trackers to heavily promote percent of peak metrics that helped clearly identify and locally contextualize the growth of second waves of the pandemic during the autumn and winter. We’re also the only newsroom using a wide array of simple algorithms to survey data from 213 countries and highlight statistically significant developments in all of them in the dynamic headlines generated on the homepage of the project. Behind the scenes, a complete readout of those headlines and custom trend projections helps direct our newsroom’s coverage of the pandemic, and can be controlled when we know we’re getting bad or incomplete data from a primary source.


Beyond the technologies we benefited from, we created some of our own, too, and open sourced a few tricks that helped us handle sticky parts of the project like translation.


Behind the public pages, we used new static site building tools to generate hundreds more individual embeddable charts from the project, which our media partners have used to inform their audiences around the world, connected directly to our live data feeds.

What was the hardest part of this project?

There are dozens of intractable problems that had to be overcome in the production of this project, but this was our team’s moonshot during the pandemic and the scale of the effort given the size of the group developing it may be what’s most remarkable.


The dev team was just three people. They designed and built the data visualizations, developed the statistical algorithms and smart texts that detect and explain critical daily news, led translation of the project in 5 languages, built and tested the backend pipelines and scrapers that manage a near constant feed of data and helped direct our COVID data team’s work and troubleshoot frequent reporting issues around the world.


This was also just one project during that small team’s year, and they continued reporting on critical news, COVID-related and otherwise, while maintaining the project and making sure it continued kicking over 24-hours every day.

What can others learn from this project?

The pandemic is the biggest story of 2020 and newsrooms with larger graphics and development teams have clearly invested a great deal in tracking it to great effect. We like to think, though, that our project makes the case for what newsrooms with fewer resources but just as strong ambitions can accomplish using open source technologies and by making smart decisions on how to scale their efforts.


From a technical point of view, that meant establishing new development patterns for complex data pipelines; using cheaper, lower-maintenance technologies to build an immense news application with over 1,000 individual live-updating pages for our readers and media clients; and writing the questions we brought to the data as journalists into simple, auditable algorithms that could detect their answers and translate them quickly for readers in all the languages we produced.


From a project management perspective, our team was spread across countries and continents. The coordination between the core dev team (in 3 different countries), the data collection teams both in-house and at third party organizations, and the production and editorial teams at Reuters was full of practical lessons for newsdev and graphics teams attempting a project at this scale.

Project links: