Unit:
Κατεύθυνση Πληροφορική της ΥγείαςLibrary of the School of Health Sciences
Supervisors info:
Φλώρα Μαλαματένιου, Καθηγήτρια, Τμήμα Νοσηλευτικής, ΕΚΠΑ
Βασιλική Κούφη, ΕΔΙΠ, Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών, Πανεπιστήμιο Πειραιά
Εμμανουήλ Ζούλιας, ΕΔΙΠ, Τμήμα Νοσηλευτικής, ΕΚΠΑ
Original Title:
Εφαρμογή για την εξόρυξη (ανάκτηση και κατηγοριοποίηση) δεδομένων απο το κοινωνικό δίκτυο Twitter αναφορικά με την ψυχολογία του πληθυσμού στην πανδημία Covid-19
Translated title:
Mining Twitter data to understand users' attitude to Covid-19
Summary:
This thesis, prepared in the framework of the InterUniversity Postgraduate Program "Health Care Management – Health Informatics" discusses Greek speaking users’ attitude with regards to Long Covid. Initially there is a mention of big data in healthcare and natural language processing. Afterwards, there is a list of use cases where natural language processing provided important insights for public health over the pandemic of Covid-19. Then the main use case of this thesis is presented. Twitter data in the Greek language referring to Long Covid were mined, processed and analysed to model main topics of discussion and analyse sentiment. Pre-processing, analysis and results were performed in Python using Jupyter Notebook tool. Results highlighted the following discussion topics: Greek-speaking users discuss Long Covid effects, Long Covid effects in specific population groups like children and vaccines. 59% of analysed tweets conveyed a negative sentiment while the rest had positive or neutral sentiment.
Main subject category:
Health Sciences
Keywords:
Big data, Twitter, Natural language processing, Sentiment analysis, Covid-19