“…We apply advances from natural language processing on the collected textual data, the method called topic modelling (TM), that has been developed in the last two decades thanks to the rapid development of computer technology and machine learning algorithms (Albalawi et al, 2020;Bing et al, 2020). 1 TM has been applied to text classification in a wide range of areas including patent data (Venugopalan & Rai, 2015;Chen et al, 2017;Savin et al, 2022c) and scientific publications (Chen et al, 2020;Savin & Teplyakov, 2022), political debates and petitions (Hagen, 2018;Wei et al, 2020), survey open-questions (Tvinnereim and Fløttum, 2015;Tvinnereim et al, 2017Tvinnereim et al, , 2021Liu et al, 2021;, firm descriptions in business platforms (Savin et al, 2022a;Żbikowski and Antosiuk, 2021) and publications in mass media (Lenz & Winker, 2020;Park et al, 2016).…”