The Corruption Perceptions Index (CPI) ranks countries and territories by their perceived levels of public-sector corruption, as assessed by experts and business people. Published annually by Transparency International since 1995, the index draws on different assessments and business opinion surveys carried out by independent and reputable institutions. This dataset covers 1995–2017.
Data sourced from Transparency International.
The CPI ranks countries/territories in terms of the degree to which corruption is perceived to exist among public officials and politicians. It draws on different assessments and business opinion surveys carried out by independent and reputable institutions. It captures information about the administrative and political aspects of corruption.
Scores from 1995–2011 are on a 0–10 scale. From 2012 onwards, Transparency International switched to a 0–100 scale; this dataset normalises those values back to 0–10 (by dividing by 10) so all years are comparable.
A score of 0.0 means the country or territory was not included in the CPI for that year. This is not a real score — the minimum real CPI score across all years is approximately 1.0.
Country names are taken verbatim from the original source files for each year. The same country may appear under slightly different spellings across years (for example, Côte d'Ivoire appears as several variants). Users should normalise country names before doing cross-year analysis.
The data in data/ is generated by script/corruption-perceptions-index-dataflows.py.
Requirements:
- Python 3
pip install dataflows pandas requests beautifulsoup4 PyPDF2 numpy openpyxl
Run:
cd script
python corruption-perceptions-index-dataflows.pyThe script scrapes annual CPI tables from the Transparency International website (1998–2009), downloads Excel files (2010–2016, 2013), and reads from local PDF and XLSX files in archiv/ (1995–1997, 2012, 2017).
Note: Several upstream URLs used by the script (files.transparency.org) return 404 as of 2024. The data in data/cpi.csv is a static snapshot last generated with data through 2017.
This dataset is made available under the Creative Commons Attribution-NoDerivatives 4.0 International license, consistent with Transparency International's terms of use for the CPI data.