Written by Gioia Deucher, swissnex Brazil & Daniel Grünenfelder, Embassy of Switzerland in Brazil
Since research about Big Data in Brazil did not yield many results, we decided to interview a university professor (Dr. Geraldo Xexéo, Prof. for Engineering and Computer Science, UFRJ) and several members of the Rio-branch of SERPRO, the biggest government-owned corporation of IT services in Brazil. SERPRO is linked to the Ministry of Finance of Brazil, developing software and services for increased control and transparency about government revenues and spending.
SERPRO administers about four petabytes of data in Rio de Janeiro alone, mostly related to the treasury department, salaries, and pensions, allocating raw data to different ministries. Big Data was situated as one of the four columns of an ‘unknown space’: 1. The Cloud, 2. Big Data, 3. Mobility, and 4. Social Networks, and opinions about its accessibility are equally diverse as the Brazilian population itself.
In Brazil, the will to share data is considered more of a cultural-dependent process, coined by individualism rather than a democratic sharing system. The ‘lei de acesso á informação’ (LAI, lei nº 12.527) –law of access to information, a result of the Open Government Partnership– states norms and regulates the release of Big Data.
While one would expect an interest of the Brazilian government in allowing access to data and sharing it, it was raised that the government seems somewhat concerned to keep data about Brazilian companies or citizens within Brazilian territory, especially since the Snowden scandal, while Big Data in the private sector is mainly used by foreign companies. Furthermore, the government reserves itself the right to implement laws and court decisions for companies to follow.
A general echo regarding open data policy was that there seems to be a (ministry-dependent) will of the government to share data locally, but that its limitations rather arise from the underdeveloped technologies and the missing credibility and knowhow about data administration. Also, a single rule could not fit all data sets, especially when it comes to privacy issues and questions of ownership. It is favored that private data remain private and government data are open, assuming it offends no privacy rights. Some of these efforts are visible in the openness of allowing access to government expenses and salaries, including making salary data comprising names, e.g., the ones of the executive branch, publicly available on the internet. Two bigger initiatives in Brazil are supposed to further engage in privacy questions over the next few years. But the fact that myriad different actors are involved, and that needs for the private and public sectors may differ, render the problem highly complex.
An initiative for centralization of data does exist in Brasília, but control of Big Data is challenging, especially when it comes to actualizations, since a good share of data is administered automatically instead of through a human being. Computers seem not to be apt for quality control yet, and certain patterns are only detectable through the human eye, especially metadata. Crowdsourcing seems to be the gold standard for detection of patterns, as they seem as unpredictable as the conjugation of irregular verbs in Portuguese. Most discoveries and applications of data seem to happen serendipitously, without any concrete intention or plan of action.
In comparison to other countries, Brazil considers itself as underdeveloped and behind in dealing with Big Data. Often, experts from abroad need to be hired, entailing high costs for companies; universities have so far only sparsely offered educational training. SERPRO mentioned a competition of ‘interpreters’ or applications of Big Data, but this seems to be very little known and so far rather an insider tip. However, there was an unanimous agreement that there is a tremendous economic potential of Big Data in Brazil, that Brazil is aware of that potential, and that major investments and efforts need to be made straightaway.