Emotions in Polish and Lithuanian Social Media

Teams of trained, native coders in Poland and Lithuania independently annotated 3,659 Polish and 1,946 Lithuanian Facebook posts (in-language and in-country) including all multimedia content but not the comments. These data were sampled from a larger dataset pulled from specific Polish and Lithuanian sociopolitical entities from 2015-2020. These annotations were on a 0 (none) to 100 (most frequent/intense) for each 23 emotions (and Positive and Negative Other) for each post. They were also rated for the personal reactions using the same emotion scheme for each post, but that data are not shared here. The annotators also coded each post for media type (e.g., text and what type vs. not, image vs. not), language, and primary and secondary topic, with topic coded using an adapted and expanded version of the Comparative Agenda’s Project’s scheme. These independent annotations went through a consensus process, and only the consensus numbers are deposited here. More detail about the sampling, consensus process, and other methodological details are available via emailing [e-mail] for draft papers. This corpus also includes the number of Facebook shares and different Facebook reactions along with other useful information (e.g., name of account, link to post).