Public Procurement Data Dictionary

The Public Procurement Data Dictionary includes the following components:

• flattened_csv – a consolidated version of all data fields merged into a single CSV file. This file can be used to perform comprehensive analyses.
• tender – variables related to the tender level, including basic information such as the announcement date, procedure type, and source URL.
• lot – all variables describing the lot level. By default, this serves as the primary unit of observation in the flattened_csv, with each lot corresponding to a contract.
• bid – information on all bidders, including details about winning and losing bids. When filtering to include only winning bids in the flattened_csv, the dataset represents a contract-level observation, where the winning bid equals the contract price and bidder attributes describe the supplier.
• buyer – variables describing the procuring agency responsible for the tender.
• unit – fields providing more granular data on individual items within each lot, when such detail is available. In countries where this level of granularity is reported, items can serve as the unit of observation in the flattened_csv, though this level of detail is not always present.
• amendment – data fields capturing any contract or tender modifications.
• funding – variables describing the source(s) of funding for a tender, where applicable.
• enumeration – an overview of categorical variables and their definitions. Currently, it is largely based on categories used in ProAct datasets.

Each tab contains a unique identifier (id) that enables merging across datasets, depending on the appropriate level of observation. For example, items should be merged by tender and lot, while buyers only require the tender_id, since each tender has a single procuring agency responsible for all its lots.

You can also download the full Excel file here.