(original post in Portuguese here)

Introduction

In the last weeks I focused my efforts on adding vaccination facilities in Fortaleza/Brazil. At the moment of writing, there are in Fortaleza exactly 113 vaccination sites, being (if I’m not mistaken) 100 clinics and 13 other places (such as shopping malls and huge infrastructures like a stadium and convention center).

The method I used for this was to use the open data of the Fortaleza City Council (don’t worry, dear warrior of licenses, we have explicit permission for this) for the locations and details of the clinics, along with the other 13 locations that have been manually updated.

Methodology

Preliminary Step

1. Initially, I obtained the scheduled lists from the Health Department website, which is nothing more than a PDF document with the scheduled people, their vaccination locations and time. I saved this PDF and converted it to Excel using Adobe Acrobat Pro.

2. Within Excel, I filtered the unique values to obtain the vaccination sites.

3. With the 113 locations separated, I split the data into clinics and the other locations. Regarding these other locations, I manually went to the OSM (using the iD editor) and added the vaccination=covid19 tag, as it can be seen in this Convention Center. Some of these places already had the COVID-19 tag, as they were the only places open in the initial phase of vaccination (around 5 places). Also, all these locations were already mapped in OSM, so the task was pretty quick and easy.

Data Handling

With these initial locations already mapped, all that remained was to identify the clinics from OSM as vaccination facilities and add the vaccination=covid19 and healthcare:speciality=vaccination tags. The problem is that very few clinics were mapped already. So, given I was willing to work, I also planned to import the clinics and perform a single task.

1. To do this, I went to the Fortaleza City Council website, Fortaleza in Maps, and downloaded all healthcare facilities (clinics, hospitals, etc.) in CSV format, totaling 200 units.

2. In Excel itself, I cleaned up the unnecessary data and converted the data to the OSM standard, such as NAME to name=*. I’ve also added the corresponding tags for each healthcare site (like amenity=clinic and healthcare=clinic for clinics), in addition to the tags mentioned in the introduction of this section.

3. I had a manual job of correcting many values, such as accent marks in the names of places and streets, besides checking the operators of each place (operator=* and operator:type=*).

Import (or not)

The data already cleaned and processed seemed fine to me, but as I had already spent a lot of time cleaning and processing it, as I had done a lot of manual work, I preferred to leave small errors to be corrected when importing.

1. I separated the data on vaccination sites and the rest of the health facilities as I wanted to focus on the vaccination sites first. Therefore, I had two different CSV files.

2. For the import, I used MapRoulette to create the tasks. I created them in the same way I explained briefly in a previous diary entry.

3. With the tasks created, as can be seen here and here, I used MapRoulette and JOSM to import facility by facility.

4. Then for each facility I reviewed the information, and manually checked if the data was correct. From what I realized, the biggest errors were in the names of the streets, which were not 100% exact as the one from OSM (Rua João Silva and Rua João da Silva), so I also took the opportunity to correct the OSM data when necessary, checking the exactly correct name, using road signs and official names according to the local government.

5. When there was already a clinic in place, I only updated the information there, to not lose its history in OSM.

6. I also took the opportunity to add service roads and other details to the clinic site, when I had the time and patience. As this was not a simple information check, this took me longer than planned.

Final considerations

After few weeks I managed to import, on my own, all the vaccination facilities in Fortaleza. In a quick consultation, I believe that Fortaleza is right now the city that has the most vaccination sites mapped in the OSM, worldwide. Obviously there is the population factor (Fortaleza has 2.7 million inhabitants), but it should also be noted that there are several German, Italian, Dutch and Belgian cities well covered.

In the rest of the world, I have seen some efforts in Russia (mainly in Ulianovsk and Moscow), Kazakhstan (Atyrau), Algeria and Romania (district of Vâlcea). In addition, the lack of mapped vaccination sites in China (due to their efforts to fight Coronavirus) and in the USA (due to their large community) called my attention.

Now I’m going to start adding/improving the other healthcare facilities remaining on my list!

External References

• One of the clinics I mapped
• Extremely practical visualization of vaccination locations, created by HeiGIT
• Types of health equipment on the Wiki
• How to map everything related to COVID-19