The Reuters vaccine tracker monitored both the progress of the vaccination campaigns in countries around the world as well as the policy schemes that decided who was given priority access when in each local health jurisdiction at a level of detail not matched elsewhere.
The tracker page was a vital resource for our readers, both to qualify how far countries had gotten in their vaccination drive and to catalogue the policy decisions being made around who was given priority access, ultimately answering the question of whether readers, themselves, were eligible to receive a vaccine.
In addition to the data feeding our own page, Reuters partnered with Amazon developers to make the data available to Alexa users. In several markets including the United States and the United Kingdom, Alexa users could ask questions and be given answers about whether they were eligible to be vaccinated or when their local health jurisdiction planned for a particular group to be given access.
Vaccine eligibility data was entered into a spreadsheet by each priority group specified by each health jurisdiction rollout plan. At a set time each day, a node process was run in GitHub Actions to scrape the data, and then a series of checks was done to both make sure errors weren’t introduced and to also flag when data hadn’t been updated within 7 days for any priority group. That feedback was sent to a Microsoft Teams channel for clear visibility to a data collection team. Once all checks had passed, the data was published to an API that drove graphics and tables on the tracker page and which Amazon developers would in turn fetch into their own ETL pipelines for the Amazon Alexa skill.
The overall vaccination statistics were visualized via a draggable canvas-based globe, which made it easy to see the yawning gaps regionally between haves and have-nots and made for an intuitive interface to scan progress around the world.
What was the hardest part of this project?
Designing a database to catalogue the disparate vaccine rollout policies of 80 countries and 50 U.S. states was incredibly difficult. Reuters needed a schema that could correctly encode the progress of each campaign in a way that was comparable across countries, generalise categories of how groups were being prioritised, capture the vaccines each group was eligible to receive and track down the dates access was promised or finally granted. Maintaining that schema as those policies quickly changed was especially difficult, and in many cases we needed to reconcile the previous plans to the current ones through our own reporting.
What can others learn from this project?
We felt like the summary statistics of how many shots had gone into arms, while important to make larger comparisons of broad policy decisions and outcomes, was actually second to a more fundamental information need our readers had in the first months of the rollout: Answering when they can get a shot. That information was harder to get and standardise and ultimately didn’t lend itself to complex graphics, but we felt it was where we were best placed to fulfil an immediate public service need. That focus paid dividends for us, and we think made ours a unique resource in a crowded field of shot counters.