Measuring the quality of large-scale automated classification systems applied to online job advertisement data

Measuring the quality of large-scale automated classification systems applied to online job advertisement data

Date: Tuesday 05 March 2024, 1.00PM
Location: Online
Online
Other Event (non-RSS)


Share this event

Measuring the quality – a quality assessment - in general- provides insight into the quality and usability of web-scraped data for producing official statistics. The quality assessment in this presentation is about Online Job Advertisements (OJA). The OJA data was assessed in a couple of ways by members of the WIN project. Firstly, the quality and stability of the sources were assessed with the help of quality indicators that were formulated for this task. Secondly, the accuracy of the classification of the OJA data was assessed through annotation exercises. We present the quality indicators and how they are operationalised in a reproducible manner for several countries. 

This webinar concentrates on using web scraping to extract data from real estate web portals for Official Statistics. Participants will be introduced to the structure of real estate portals in Germany and the importance of data from the web for Official Statistics.

For more information or to book your place visit our Eventbrite page.