Posters

Insights from Twitter text mining in ophthalmology

Poster Details

First Author: M.Mahabir INDIA

Co Author(s):    N. Shree                    

Abstract Details

Purpose:

There has been widespread penetration of internet services. Social media has become a ubiquitous part of our lives. Analysis of social media like twitter gives an insight into the dynamic trends, needs, sentiments and aspirations of the users. It has been widely used in many industries, including healthcare. This study aims to find the most commonly used terms in tweets by the users including common people, journals, medical societies and diseases.

Setting:

Rajendra Prasad Centre for Ophthalmic Sciences, All India Institute of Medical Sciences, New Delhi, India

Methods:

Keywords including twitter handles relevant to ophthalmology were identified. Some of the search terms used are: ESCRSofficial, aao_ophth, jamaophth, BMJ_ophth, ophthalmology, cataract, lasik. R statistical software was used for this study. �â�€�˜twitteR�â�€�™ package was used for retrieving the tweets, �â�€�˜tm�â�€�™ package to clean the messages. The search term itself was removed from the messages. 'wordcloud' and 'RColorBrewer' package was used for generating a colourful wordcloud visualization of words. Most frequent words were in the centre of the cloud with larger font size. A bar chart was generated showing the 10 most frequently used associated terms.

Results:

Various search terms and most common associated relevant word in decreasing order are as follows (as on 15th of March 2017):ESCRSofficial: ophthalmology, escrs, thanks, daily, drfishman.aao_ophth: glaucoma, ophthalmology, care.Jamaophth: glaucoma, injections, risk, surgery, repeated.BMJ_ophth: ophthalmology, british, journal, macular. Ophthalmology: patients, stem, cell, treatment.Cataract: pls, surgery, eye.lasik: eye, surgery, best, laser.The results are like snapshots in time are likely to change if the search is done at different times highlighting the then prevalent sentiments and trends.

Conclusions:

The most commonly used terms have been reported by this study. While some of the results might be self-evident, other results might require an in-depth look into the context in which the terms have been used. They give insight into the areas of high interest as well as biases. They also highlight the heavy influencers. Sophisticated temporal and geospatial analysis of the social media like twitter gives insight into various aspects and may help in course correction and better planning. Insights from social media analytics may bring the caregivers and common people closer together in their hopes aspirations and expectations.

Financial Disclosure:

NONE