WIN project
The State Data Agency (Statistics Lithuania), together with 17 other organizations from 14 European countries, participates in the „Trusted Smart Statistics – Web Intelligence Network” project. The project is a continuation of previous activities carried out in 2016–2020 (the “ESSnet Big Data” and “ESSnet on Big Data II projects), which targeted multi–purpose statistics based on an array of non–traditional data sources. The key goal of this project is to contribute to the establishment of the Web Intelligence Network (WIN) across the ESS and beyond for the production of statistics incorporating data retrieved from the Internet.
Testing of the network to be developed is performed by analyzing data from two domains: online job advertisements (OJA –- Online Job Advertisements) and data read from the Internet, used to assess enterprise characteristics (OBEC – Online-based Enterprise Characteristics). List of enterprise websites is necessary for evaluating characteristics of the enterprise by using scraped data.
Compilation of the list of enterprise websites and evaluation of the indicator (percentage of enterprises with a website)
OBEC population consists of legal entities of all types and forms of ownership, the activities of which according to EVRK Rev. 2 (Classifier of types of economic activity) are assigned to sections C–J, L–N and S, with 10 or more employees. Based on administrative and private sources, some of the legal entities are assigned their websites. For the rest of the population, the assignment of websites (or the indication that the surveyed legal entity does not have a website) is carried out using a URL (Uniform Resource Locator) search program. The latter consists of several stages:
searching for websites of potential legal entities by entering keywords into the Internet search platform;
scanning websites of the potential legal entities found and finding the necessary indicator values;
classification of websites of the potential legal entities found into suitable (the website under consideration is the website of a specific legal unit) and unsuitable (the website under consideration is not the website of a specific legal unit).
In the first stage, name of the legal unit, code and municipality where the legal unit is registered are selected as keywords. In the second stage, 8 indicators are selected: “does the website in question contain the code of a specific legal entity”, “does the website in question contain the address of a specific legal entity”, “does the website in question contain the telephone number of a specific legal entity”, etc. Both deterministic and machine learning methods are used to classify websites.
The result obtained – the available list of legal entities is supplemented with an indication of whether a legal entity has a website or not. Since 2022, data have been published by statistical unit – an enterprise (a separate legal unit or the smallest organizational combination of legal entities that produces goods or provides services, has a certain autonomy in making decisions, distributing its existing resources, carrying out one or more types of activities in one or more places). The resulting list of legal entities is transformed into a list of enterprises by applying the following rule: if at least one legal entity belonging to the enterprise has a website, then that enterprise has a website. The percentage of enterprises with a website is assessed by type of activity, region, county, municipality (see below).
Assessment of other enterprise characteristics, such as “does the enterprise’s website have a link to social networks”, “does the company conduct e-commerce”, etc., is being tested and will be published later.
Data compiled: November 2023.
Contact details
Vilma Nekrašaitė-Liegė
El. paštas: [email protected]
Tel.: +37065040807
Explanations of indicators:
Explanations of indicators:
Explanations of indicators:
Explanations of indicators:
Explanations of indicators:
Explanations of indicators: