"National Insurance Information System" and HFLabs launched a project to improve the quality of data

2024: Improving Data Quality

The National Insurance Information System and HFLabs have launched a project to improve data quality. HFLabs announced this on April 15, 2024.

Improving the quality of data contained in the automated information system (AIS) of insurance is one of the priority areas of NSIS activity. As an operator of a single AIS insurance, NSIS received a large array of accumulated historical data from operators of various information systems. While the data uploaded to the insurance AIS is already subject to strict control at the entrance - including address data for compliance with the structure of the State Address Directory (GAR), which is issued by the Federal Tax Service, historical data are very heterogeneous in quality.

During the pilot project, NSIS and HFLabs, using the Factor software developed by HFLab, analyzed about 5 million records on home insurance contracts in which addresses were transferred by text line. The task of the "Factor" was to analyze them and structure them by components (country, city, street, house, etc.) in accordance with the GAR. As a result of the work of the "Factor," it was possible to successfully restore 88% of the lines to the full address by GAR.

With the help of the "Factor," the NSIS will be able to determine correct and incorrect addresses and, if the data is not valid, then return them to the insurance company for adjustment. In the future, the Factor will be able to enrich addresses with GPS coordinates if the insurer does not indicate them. Factor will also help automate processes for processing large amounts of address data - for example, when preparing mailings, - said Nikita Nazarov, technical director of HFLabs.

By law, the AIS insurance operator is obliged to provide authorized bodies information with about insured objects real estate in a certain area (for example, in the area of ​ ​ flooding). It is impossible to implement such functionality without an exact location address and coordinates of the structure. Correct address determination is also critical for other types of insurance - for example, an error in an address can distort the values ​ ​ of tariff coefficients in contracts. Using MTPL specialized software for address recognition, such as "Factor," NSIS can not only control the correct direction of data by insurance companies, but also recognize the addresses indicated in scanned copies of documents provided by insurance companies or insurers, facilitating the client path when concluding and maintaining insurance contracts, "said Olga Mordkovich, Deputy General Director of NSIS.

As part of the pilot project, NSIS also tested the client MDM (data management systems) from HFLabs - the Single Client product. AIS insurance uses and plans to develop its own MDM, but the study of HFLabs approaches and possibly the use of separate components "Single Client" can help the AIS insurance operator process large amounts of personal data that appear in different types of insurance contracts.