A collaborative effort between Público and other 15 European news organizations, coordinated by OCCRP, found out that the European market was being flooded with fake and faulty FFP masks.
Using a data-driven approach, I’ve found out that Portugal has bought at least 627 thousands of FFP masks that had a fake or dubious certification. Some certification companies that were not allowed to certificate FFP masks started to issue documents that looked like certificates and fooled public institutions that were not used to buy this kind of product.
This international investigation made local authorities act across Europe. In Sweden, the local authorities demanded masks delivered to several hospitals being returned. In North Macedonia, the Health Ministry called for an investigation.
But things happened a bit differently in Portugal – a country where, the data collected showed that at least hospitals, the Marine, and even the Health Authorities bought some masks with dubious certification. After several questions, the Ministry of Health didn’t confirm if there was an investigation happening because of these public contracts.
According to the Attorney General’s Office, only one company that sold this kind of equipment was being investigated.
This investigation was born while I was working on another investigation about public spending during the covid-19 pandemic. I’ve used R to scrape the Public Tenders website from Portugal to get all pandemic related contracts. After going through the description and pdf contracts of about 16 thousand contracts, I’ve identified 232 public contracts that specified purchasing FFP masks – I’ve used NLP and machine learning approaches here to help me identify possible FFP contracts.
Because most of those 232 didn’t provide additional information on the public website, our team called all buyers and all sellers asking for the certification of the masks bought/sold. Most of them refused to give the documents and, by the date of publishing the investigation, we had 50 certificates of those masks.
While doing it, across Europe other journalists were doing the same and checking with experts in the area the validity of those documents so that we could end up with an international database of fake/dubious certificates. That way, we found out that more than half of the certificates provided to us from public contracts provided this kind of dubious certification.
What was the hardest part of this project?
The hardest part of this investigation was navigating the sea of data. Since we had around 16.000 public contracts that were related to the pandemic, finding which ones were about FFP masks was hard – especially because most of the contracts were vague and simply stated that “masks” were bought (not specifying the kind of masks). Automation and data journalism techniques helped here, but when there are missing data, nothing works better than using the phone. But calling all the buyers and sellers asking for a document was hard and involved a lot of people and a lot of hours by the phone.
What can others learn from this project?
This is probably something that all investigative journalists say, but since this was my first time taking part in an international investigation, it was the first time I felt it: international collaboration is key to tackle international issues.
Without the collaboration of the other 15 newsrooms (and OCCRP) it would be impossible for a single data journalist in Portugal to know about this issue, to have some background about the companies that were issuing these certificates, and to be able to check if a new certificate was ok or not.