Dozens of requests for access to public information made by Salud con lupa and the crossing of databases to get close to the situation of Amazonian indigenous peoples during the pandemic.
The Other Peru shows the role that data analysis can play in detailed coverage. It demonstrates how transparency mechanisms can help close the information gaps that otherwise increase the isolation suffered by native populations. The project built and analyzed databases containing information from various reliable sources to prepare a detailed study of the situation facing native communities in five regions: Loreto, San Martín, Madre de Dios, Ucayali, and Amazonas.
Tracking the impact of COVID-19 in Peru’s remote Amazonian jungle with spotty cell phone service is a daunting task. But this is precisely what Salud con lupa did for months to make happen. Through on-the-ground reporting, our team found that the number of fresh graves in just one community was three times higher than the number of deaths the government reported for the entire region.
Little access to diagnostic tests, a spike in dengue fever, and a reliance on natural medicine were among the reasons why many COVID-19 deaths in the Peruvian Amazonia were left out of the official count. The most important reason, however, was “a lack of care.”
With maps, statistical graphics, and other visual representations, the platform provides information in various layers to help understand the conditions in which native communities are living as the pandemic unfolds. This is a topic mostly overlooked by mainstream media, whose coverage tends to have an urban bias.
At the first layer, the user can find—by district—the location of population centers inhabited by various Amazon ethnic groups. This mapped data also shows the location of the several types of health facilities in each district and also demographic, economic, and social data, such as access to basic services.
At the second layer, the platform enables the user to monitor the progress of the Covid-19 vaccination program in the native populations of each district. At the third level, the platform displays the incidence of endemic infectious diseases in the Peruvian Amazon: Zika, dengue, malaria, and chikungunya. Finally, it breaks down, by indigenous group and by district, the number of infections and deaths due to the pandemic.
Using OpenRefine software, we cleaned up and standardized the databases, a process that allowed us to save the information in CSV format files and facilitated its tabulation.
We relied on the JupyterLab platform—which uses the Python programming language—to process, analyze, and cross check the databases, making it possible to create sequences to automate information processing. We also used the Pandas, Numpy, and Matplotlib libraries together with JupyterLab for data exploration and transformation. As is well known, a library is a set of scripts or lines of code that can be reused to perform specific tasks.
The result is a powerful and user-friendly tool that contributes to understanding the challenges indigenous communities face today in the Peruvian Amazon. This tool—which we have made public together with various other pieces of journalism—is a contribution not only to the work of researchers interested in this often-forgotten sector of the country’s population but, above all, to decision-making. With more precise and complete information, better public policies are possible.
What was the hardest part of this project?
The principal challenge was the requests for information from authorities. We were constantly told that there is no data, or only partial data. Most of our attempts were successful if we contrasted the official numbers with the information that people gathered in the field. That information was collected by interviewing many indigenous leaders as a more reliable source.
Peru’s Minister of Health reported 148 indigenous deaths during the pandemic. This is false. We know that more than 400 died, just in one community. It shows you that they are not taking care of the data, and they are not documenting what happened. It was our goal to tell the real story.
The work began with data collection, made difficult by a lack of information from official sources because health authorities failed to build ethnic identity into their approach. Although they later corrected this oversight, they did not then make the associated data freely available to the public.
On vaccination, the ministry reported that it did not have a specific plan for indigenous communities. On other matters, it suggested consulting regional authorities or reviewing the so-called Indigenous Situation Room. The latter is a web platform created by the national government that presents snapshots of Covid-19 case and death numbers by community, ethnic group, and geographic district. However, the platform did not allow downloading of raw data, and, in some instances, gave incomplete information.
We sent requests for access to information to the regional health directorates (DIRESA). We attempted to contact the directorates of the five priority regions, but once again encountered barriers. The government of Amazonas did take our request but then never responded. Only two regions provided information on the number of infections and deaths, having had the foresight to record ethnic groups in their Covid-19 statistics.
What can others learn from this project?
The project exemplifies a challenge in countries like Peru and other Latin American countries that don’t have enough data. In this case, we focused on indigenous communities in the Amazon.
Since the pandemic started in Peru, we didn’t have data about this population because of a lack of access to health services. Nobody was registering COVID-19 deaths. That’s why officials and the Minister of Health didn’t provide an official report about the impact of the disease in this part of the country.
We worked to gather data and to try to understand what happened. To do so, we had to make a lot of requests for information. We built a database on our landing page and featured stories to provide an explanation. It’s important to investigate because indigenous communities are at risk if we don’t identify who died.
It’s very important to work with requests for public information or public data. But if that falls short, try to build the story with other sources. Many people may read the project because they find little substance [when they] search for information.
You know that there is a reality in front of you, but it is not revealed by official numbers. Some reporters will say that if we don’t have official data, we don’t have the story, but it’s not true. There are alternative ways to approach the problem and paint an accurate picture of the situation.