We analyzed 7,114 secret decrees signed by the last Argentine military dictatorship between 1976 and 1983 that, until now, had never been fully processed and reviewed. From a quantitative and qualitative analysis, we identify orders and dispositions, from arrests and deportations, to the arms trade, editorial bans and real estate operations, through million-dollar contracts to counteract the bad image of the dictatorship abroad. Everything was decided without society knowing and was revealed in the work presented to apply for this award. This series of articles began to be published on March 24, 2019 and extended until May 26, 2019.
Between 2013 and 2015, the government of Cristina Kirchner ordered the declassification of a large number of secret decrees of the dictatorship that had not been published at the time in the Official Gazette. But this declassification was done in a disorderly way, at different stages over three years, and in a little bit visible section on the Official Gazette website. In practice, the vast majority did not meet publicly, nor did anyone analyze them in their entirety.
We carry out a deep journalistic research of historic background and political context to help the public to understand possible reasons for classification by dictatorship governments and to exposed their arbitrariness.
Through a series of requests based on the Law on Access to Public Information we found that there were almost 1,500 decrees signed by the dictators that still remained secret and that it was not known what they were, nor the reasons why they were not declassified along with the rest.
In this process, we also identified 8 decrees that had not been declassified by Cristina Kirchner and we revealed their content and the possible reason why they remained hidden.
We decided to extend the analysis to the secret decrees issued in democratic periods from 1958 and we identified, thus, that there were almost 600 norms signed by constitutional presidents that remained outside the public light and nobody knew about them
Our investigation exposed those decrees that still remained secret and the area of the Government they were referring. We identified which were in the process of being declassified by Mauricio Macri administration. Through a request of the Law on Acceso to Public Information we advanced the content of a dozen of them. The open data of all decrees were shared with society.
From the Infobae Data Unit we proceed to the complete download of the secret decrees, including 354 attached documents. After the opening of data (a PDF conversion was performed for TXT), the documents were uploaded to the Document Cloud platform, which allows access to them in the original format and in text format. We used an OCR system (optical recognition characters) to open the data. Then, the counting of the decrees for the time series was automated. Data parsing was performed using Python. Repeat patterns were relieved, through different variables that were put under study. In order to visualize the volume of secret decrees during the dictatorship, in comparison to other periods in Argentine history, we used Flourish. The open data of the secret decrees in text format were shared with the audience for the first time, through a Google Drive link for public consultation.
What was the hardest part of this project?
In order to process and analysed the 7,114 secret decrees into a series of journalistic notes, the Infobae Data Unit proceeded to the complete download of all of them, including its 354 attached documents.The most difficult part was the reconstruction of loans taken by other nations for the purchase of military weapons. The design of the spreadsheet had to be done manually, to reconstruct the timeline of money and the handling of figures in different currencies. Then perform the calculations of the value of money to report what value in dollars the indebtedness would have at current figures. As it was not just to analyze only numbers, but text, another difficult part was to detect repetition parameters on words. First, a sample of 400 random decrees had to be read to identify them.
What can others learn from this project?
The team that worked on this project was composed of Mariel Fitz Patrick as a reporter, Daniela Czibener as a programmer, Ezequiel Di Martiis as DataViz and Sandra Crucianelli as an analyst.
There was was not only a text-based data analysis but also journalistic research of background and context. The opening task combined digital and manual work, through information requests and manual data entry. We only use free tools that could be used by any other reporter. We were only moved by the desire to open information, process it and expose it in the light of public opinion.