InBillo - data for the model allowing for the assessment of the economic situation of the market participant

Within the context of the POIR.01.01.01-00-0247/20 initiative, financially supported by the Fast Track 3_2020 competition for the Mazovia region under the theme "Service for securing and releasing financial liquidity for companies - Inbillo", an automatic scoring system was implemented to assess the economic condition of market companies. The developed scoring method enables the use of information about companies available on the internet to assess business risk.

Between August 1 and December 31, 2021, over 22,376 analyses assessing companies were performed.

Tests showed that the system autonomously classified 10% of companies as unauthorized to participate in the market, assigning them the status of "untrusted". This is considered an exceptional result, based on the analysis of data such as:

• _id - database identifier + NIP - taxpayer identification number
• refer_scoremain - Main RS indicator
• refer_scorecustomer_trust - RS customer trust indicator
• refer_scoredevelopment_advance - RS technological progress indicator
• refer_scoreorganization_maturity - RS organizational maturity indicator
• refer_scorepayment_morality - RS payment morality indicator
• wwwdomain - website address
• activity_company - company activity status (Active)
• company_dataemployment - employment information
• company_datalegal_form - legal form of the company
• company_dataestablishment_date - company establishment date
• debts_mr - debt information
• subsidies_sudoptotal_benefits - SUDOP benefits
• whiteliststatusVat - VAT white list status
• wwwsocialmedia_list - social media list
• wwwtechnologies_list - technology list The algorithm revealed that companies deemed untrustworthy typically had low scores in areas of organizational activity, payment morality, and customer reputation. Particularly, the latter aspect is crucial in identifying unauthorized market participants. Internet users often publish alarming information about companies, such as marking domains as fraudsters in NASK, or negative reviews on Opineo, Google Maps, Gowork, which prompted the system to search the internet in real-time and learn the characteristics of companies that could pose a future threat to the market.

The attached dataset serves as a reference material for analysts, enabling the training of their own scoring systems in accordance with the credit policies of their organizations.