Procurement Analytics for Kazakhstan and Uzbekistan
Analysis of state and quasi-state procurement: item classification, matching of repeated procedures, and search for buyers and suppliers. Notices, award protocols and contracts are linked to assess demand and the competitive landscape.
The task
Build a procurement analytics system for Kazakhstan and Uzbekistan. Bring together data from different platforms for market research, competitor analysis and the selection of tenders matching a supplier’s profile.
Data and domain logic
Notices, specifications, award protocols and contracts reflect different stages of a procurement. Repeated procedures and document revisions may relate to the same buyer requirement.
The system links these records and reconstructs the history of a lot from publication to outcome. This makes it possible to assess demand without counting the same procurement twice.
Analytical capabilities
Item classification using machine learning and language models, analysis of buyers and winners, and determination of procedure outcomes are under development.
Procurements tied to an installed platform or a specific rights holder are treated separately. The analytics help distinguish competitive procurement, continuation with an incumbent supplier, and unsuccessful procurement procedures.
Status
In development. Source connections and the rules for processing data from the local platforms are being set up.