• 1
Wednesday, March 29, 2023 at 06:30 p.m.

Dementia and anxiety take center stage in citizen science conversations on Twitter

A study carried out by the URJC has analyzed the topics on health and healthcare available on this social network to find out its impact and scope. The results reveal that there is an active conversation about different diseases and cognitive disorders.

Irene Vega

Through the analysis of citizen science and health conversations on Twitter, a research team from the URJC, in collaboration with the RIAS Institute, has discovered that one of the main concerns of users (beyond COVID-19) is the issue of mental illness. In addition, most of these messages revolve around the Sustainable Development Goals (SDGs), being especially related to SDG 3 Health and well-being.

The results of this workpublished in the scientific journal Digital Health, reveal that there is an active conversation about various diseases and disorders, especially dementia and anxiety, followed by COVID-19, diabetes and cancer. "Interestingly, the results found on COVID are lower than expected after the pandemic," highlights Fernando Martínez Martínez, a researcher at the Higher Technical School of Computer Engineering (ETSII) and co-author of the study. “In addition, it notably highlights a very active conversation about the Mosquito Alerta app and a citizen science platform aimed at monitoring sightings of tiger mosquitoes and mummy mosquitoes that transmit Dengue, Zika and Chikungunya, which has shown a high participation of the citizen science community on Twitter both in the creation of content and when publishing. share this initiative”, adds the researcher.

This research represents an important value as information for the possible creation of recommendations on health policies through the use of data on trends, influential accounts or thematic. “Within the study, we confirmed that the most influential accounts with the largest number of followers and retweets they were those that belonged to organizations and projects”, points out Fernando Martínez.

In addition, thanks to the high replicability of this analysis model, its application can be used to expand the study on these topics and monitor the evolution of trends, users and other conversations. In this sense, according to the URJC researcher, "the different results also point to new topics that would be interesting for a study of their own, such as the use of apps for monitoring diseases, mental disorders or research into rare diseases, as they are topics very approached”.

Analysis of the messages (tweets) and labels (hashtags)

In order to get to know in detail what the conversation about health and healthcare was like, the scientific team created a set of keywords used to filter all the Twitter data they collected. These data were collected from the tool lyguo, which is responsible for extracting messages related to science from Twitter. "After getting those tweets related to health, we used classical social network analysis techniques to obtain the hashtags used, the evolution of its use over time and the users with the most retweets”, explains Fernando Martínez.

They then used more modern natural language processing, or NLP, techniques. Natural language processing) and NER (Named entity recognitionn) to determine which were the most used terms in a conversation. They also used graph analysis techniques to create networks on the connections between users who talked about a certain topic and, in this way, observe who were the main senders and receivers of the messages. Lastly, according to the URJC researcher, "we used the techniques of Topic Modeling y Machine Learning to analyze which were the most discussed topics and how they evolved over time, as well as which were the most frequent languages ​​in this field”.

This work is part of the CS-Track project, led by the URJC and whose objective is to analyze existing data in different media related to citizen science.