Applying short text topic models to instant messaging communication of software developers

Content related to paper "Applying short text topic models to instant messaging communication of software developers" published in the Journal of Systems and Software. Data available: Data sets used: JSONs with messages from Gitter chat rooms (downloaded with previous Gitter API - https://developer.gitter.im/docs/welcome): "Android.json" "ConsenSys.json" "WebpackDocs.json" "Jenkinsci.json" "Locomotive.json" "SpringSecurity.json" "Flutter.rar" - json file was compressed due to its size "GitterHQ.rar" - json file was compressed due to its size "Laravel.rar" - json file was compressed due to its size "stopwords_list": Customized list of stop words "topics_sttm_results.csv": Topics obtained with each combination of model and corpus (both lemmatized and stemmed corpora) "intrusion_tasks.csv": Results of the survey for the Intrusion Tasks and its participants' background "topicnaming_tasks.csv": Results of the survey for the Topic Naming Tasks and its participants' background "intrinsic_metrics.csv": Scores of topic coherence metrics at topic level ('average' represents the score at model level) "topics_themes_chatrooms.csv": Results of the exercise described in Section 5.2 with the topics and themes identified in each of the 87 Gitter chat rooms. "sensitivity_analysis": Results of a smaller-scale sensitivity analysis to check the impact of the number of topics on the main findings of the paper.