A colleague asked whether there are 'low-hanging fruits' to improve the response to the pandemic.

I know two for all those who publish original data:
– Publish data in a csv file. (Not in pdfs, press conferences, or Twitter images.)
– Provide a clear description of that data.
I know of researchers who build AI and machine learning tools so that they can read the pdfs in which countries are publishing their data.

No one should spent their time on that.

All we need is a spreadsheet + a clear description of what this data tells us about the world.
Some have the impression that this is only a problem of poor communication with the general public.
But this affects everyone, including the very best researchers in the world. They waste their time hunting down obscure pdfs in changing locations with data that is not documented.
For international organizations that compile the data on the pandemic it is the same.

Every single day they need to go through hundreds of poorly documented sources. Some health ministries publish their latest data *only* in posts on Facebook and Twitter.
Our main effort is to build a global database on testing for COVID-19.

Our page on it [ http://ourworldindata.org/covid-testing ] includes a Checklist for all the information that should be clear from the data description of the original publisher.

Below is part of that checklist.
It is possible and in everyone's interest to fix this.

If you are involved in the publication of data it would be great if you could get your team to publish a spreadsheet + clear data descriptions.

Thanks!
You can follow @MaxCRoser.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: