This project will develop a first-of-its kind platform to access real-time and up-to-date public procurement data from 26 countries. The platform will also automatically assess the data for its quality and determine it's usability for key diagnostics of integrity and anti-corruption, effectiveness, and efficiency of procurement. The data can then be used for red flags detection through both low tech and advanced means (including use of artificial intelligence and machine learning).
The project proposed has three components.
- Public Procurement Data Registry: A public platform to find public procurement data from all governments publishing open data according to the Open Contracting Data Standard (OCDS). This registry will be designed with non-technical users in mind and will build on an existing suite of open source tools.
- Open contracting data quality tool: A software that can be used to report on the quality and completeness of public contracting data sets. The software is enabled by the fact that dozens of governments are already publishing their data openly in a standardized format. The tool will help data users globally to understand whether the published data is actually usable for business intelligence, detection of corruption red flags, procurement performance diagnoses and other monitoring purposes.
- Diagnosing the efficiency, effectiveness, and integrity of public procurement systems using the resulting data. This will include key indicators or 'red flags' of corruption that can be used as a basis for policy recommendations.
For example, an IMF public financial management expert may wish analyze the effectiveness and integrity of budget execution. The tool would enable them to find out which countries are publishing data, to understand its quality and completeness, and to select indicators for analysis, such as the proportion of single bid tenders (which tend to be ~8% more expensive for public budgets).
Developing these innovative tools for finding and loading data, ensuring its standardization, quality and completeness will be foundational for use of artificial intelligence, machine learning and blockchain applications. This project will deliver the real-time, high quality, structured data, that is essential for corruption detection analysis using artificial intelligence and machine learning.
Our proposal has already generated excitement at country level. We have received great feedback on our first blog post about the data quality tool (named 'Pelican'). Both the government of Canada and Afghanistan expressed an interest in using it already. In particular, we are excited to use the tools with the government of Paraguay to develop corruption detection processes using machine learning. Aslam Khan from the National Procurement Authority of Afghanistan has officially joined our project and staff from the World Bank and the IMF have joined the team: Kristina Aquino, Johannes Kiess, and Vladyslav Rashkovan.
Our project will be useful to several other proposals in the challenge that involve the use of OCDS data and open data on public procurement more generally. Any proposal that will use data in procurement will require an understanding of the quality of data and which checks can be reliably calculated. In particular, we are working with Poder and Nosotras on open contracting in health and SIF & CoST on open contracting in infrastructure.