Project description:

The article describes the number of people currently being treated in intensive care and inpatient care as a result of Covid-19 in Sweden and its different regions, as well as additional demographic and medical statistics about the people being treated. For exaample age groups, risk groups and mortality rate. The article is automatically updated whenever new data is available. It should work ok to google translate the information. 

Impact reached:

Due to the fact that very few people were tested for Covid in Sweden in the first months of the pandemic, measuring the number of people in intensive care became the most reliable indicator of how well (or how badly) the fight against the disease was going. Through the article, we provided a way for the public to inform themselves about the current status of the outbreak nationally as well as in their own regions.The article is often referred to in discussions about how the pandemic is affecting Sweden, both in domestically and internationally, and has, as of this date, had 32,4 million page views – that is alot in our mediahouse. People come back to this page every day or a couple of times a week to understans the situation. We get at lot of emails form doctors and nurses that prefer using our page instead of other sources – that’s how we know we have done our job.

Techniques/technologies used:

In order for the data to be as up to date as possible, we set up an automatic data pipeline where a python-script on our server would collect the latest data and upload it to a file server for our visualizations to use. There was no public API we could use, so in order to get the information we needed from the official source, we had to reverse-engineer http-requests sent to their server and figure out what parameters to change in the call to get the correct data.The visualizations were made in javascript and html, with the help of d3 and Svelte. We wanted the reader to be able to easily compare trends in different regions, and allowing them to get more detailed information in specific regions, so we decided to display mini-versions of the charts in grids. Clicking on the mini-version gave the reader a more detailed view. We also built warning-bots that let’s us know if someting looks strange with the numbers at our sources webpage. This has been useful for us, but also for them several times when we have discovered faults in the data long before they do. The article is design to work equally well on desktop and on mobile.

What was the hardest part of this project?

One of the challenges of the project was creating a stable and reliable infrastructure to collect and verify new data. In the first few months especially, we had to quickly adapt to changes in the data other hiccups that would occur in the systems of the agency providing the data.We established a relationship with the people responsible for the data and tried to include as many checks and fail-safes as we could to ensure that the data would update as intended.
Another challenge has been staying on top of new information as it became available. When we first published the article there was no information about the number of people being treated in intensive care any given day (the only information we had was how many new individuals had been checked in to intensive care each day, we didn’t know how many were still being treated). We also had no information about inpatient care, and so when the new data suddenly became available, we had to scramble to quickly set up the new data pipelines and create the new visualizations.

What can others learn from this project?

Our project shows how basic chart designs can be effectively utilized to communicate critical information in a simple yet powerful way, which both allows the reader to get quick overview of the current situation, and allowing for more detailed exploration of the specifics of the data.

Project links:

www.svt.se/datajournalistik/corona-i-intensivvarden/