Mining Facebook Data of People with Rare Diseases

This research is concerned with the study of Spanish Facebook pages that deal with rare diseases. The objectives of this research are to characterise these pages and to compare them with the priorities of the Decalogue of the Spanish Federation of Rare Diseases (FEDER). This research uses Netvizz to download the data, word clouds in R to perform text mining, TextBlob in Python to perform sentiment analysis, and log-likelihood in R to compare Facebook and Decalogue words. The results obtained show that photos are the type of posts with higher number of likes, reactions and engagement. We can also see that positive polarities have higher level of engagement, and that subjectivity is not so correlated with engagement. In the comparison of the Facebook data with the FEDER Decalogue, we observe that the following words have a lot of presence in the Decalogue and little in Facebook: disability, professionals and diseases. Similarly, these are the most present on Facebook with little representation in the Decalogue: help, life, people and children. In conclusion, we can say that the Decalogue should focus more on help, life, people and children and less on disability, professionals and diseases.