An Approach for Researcher Identification on Twitter Without the Need for External Data

Background: Keeping track of the developments in a scientific field can be challenging. Techniques for automated content analysis represent a promising approach for getting insight into large text corpora. Topic modeling, in particular, is gaining in popularity in scientometrics. A topic-guided and user-friendly interface for databases of scientific literature can open publication trends to a broader audience with various user scenarios. Aim: The goal of this project was to develop a user-friendly web-based application for exploring and analyzing research topics in Psychology. Method: A standardized vocabulary of keywords in PSYNDEX - the database of psychological literature from the German-speaking countries - was the input for topic modeling based on latent Dirichlet allocation. A total of 329,240 documents published between 1980 and 2017 were included. Results: The final model comprised 213 topics. The app features different views for exploring the topics: "popular topics," "hot topics," "cold topics," and an overview of "all topics." Each topic entry has a search button that forwards a search query to PSYNDEX for literature relevant to this topic. Conclusions: Initial user experiences confirm the app's ease of use. Future developments are discussed. A demo version of the app can be accessed for free via https://abitter.shinyapps.io/psychtopics/.

Understanding and promoting researchers' well-being is crucial for achieving successful research outcomes and fostering a thriving scientific community. While conventional well-being assessments can be resource-intensive, social media language analysis offers a promising solution. This research-in-progress aims to evaluate the potential of Twitter (now “X”) data in providing insights into researchers' well-being. To address this research question, we will derive researchers’ well-being from our available Twitter dataset and use survey data on a known event (i.e., initial phase of the COVID-19 pandemic) for external validation of our results. By assessing the eligibility of data from academic social networks as a well-being indicator for researchers, we aim to lay the groundwork for real-time monitoring for informed decision-making and foster a supportive research environment.