El indultómetro: Investigating pardons granted in Spain since 1996


El indultómetro is an exhaustive data analysis and investigation on the use of governmental pardons in Spain since 1996. Through data scraping of the pardon announcements in the Official Gazette, the project classifies and visualizes all the pardons granted to date, allowing readers to quickly and easily filter pardons by type of offense, compare annual data, and check how different governments have administered the power.

Since the project’s beginnings in 2013, the number of pardons has dropped significantly: from 1.5 pardons been issued a day to 1.5 per week.

Image: Image: Pardons before and after El indultómetro.

Building the database

The project runs off data downloaded from Spain’s Official State Gazette, which publishes a record of all pardons granted in the national territory.

Once downloaded, the team used Ruby and Regular Expressions to identify and separate pardon information from unrelated information in the records. They then undertook a new data extraction process to fill the following data fields for each pardon:

  • Identifier of the Gazette provision
  • Publication date
  • Ministry: Justice or Defense
  • Gender of the pardoned
  • Sentencing court
  • Date of conviction
  • Role of the pardoned in the crime
  • Crime and sentence
  • Year of beginning and end of criminal conduct for which the individual was convicted
  • Type of pardon: commutation of sentence, partial pardon, total pardon
  • Reduction / New Conviction
  • Conditions for pardon granted
  • The date of granting the pardon
  • Signing minister

Because much of this work was based on complex text processing, there was a risk of data errors. Therefore, the team implemented a double process to verify their results. Firstly, they undertook an internal verification process to manually debug pardons that were unable to be programmed automatically. As a second verification measure, the team used the Gazette search engine to search for all mentions of the word ‘pardon’ on the site. By comparing results in an Excel spreadsheet, they were able to identify some pardons that were missed due to misspellings in the title of the official record.

Once the final inventory of pardons was cleaned, the team categorised different crimes and offences for which pardons had been issued using the classification present in Spain’s Penal Code. To simply the process, they ruled out nuances, like where a crime was aggregated or tentative, to minimise the breadth of different crimes to be classified. Even so, the team ended up carrying out a manual classification of over 300 crimes and referring to content specialists for review. 

Image: The classification process laid the foundation for the database's crime category filter.

The project’s data is available to explore online, and there is also an API where you can access the database for your own analysis.

The results

Building upon their database, the team conducted several investigations that found:

  • Crimes against the environment, crimes committed by officials against individual freedom, prevarication of public officials and embezzlement add up to the highest percentages of grace measures for convictions
  • Of the 20 fastest pardons granted between 1996 and 2013, 10 were convicted of embezzlement of public funds, kidnapping and or illegal detention related to GAL.
  • 227 pardons have been granted to those convicted of corruption

Image: Pardons for corruption crimes by different governments.

Explore El indultómetro here.

Parts of this post were originally published in Spanish, available on the project site. Image: Ramón Durán.