People use social media platforms like Facebook, Twitter, and blog sites for expressing their views and criticising the products purchased and movies watched. They use these platforms for getting information like blood donation requirements and job opportunities. During the disastrous situations like floods and earthquakes, these platforms act as powerful media for passing messages to all people. During this COVID-19 pandemic period, all social media platforms are effectively used by all businesses for the instant communication and interactions between the groups of people. In all these scenarios, the information gets diffused and reaches different levels of people. Sometimes this diffusion gives positive aspects to the readers; sometimes it creates negative impacts to them, which has its own cascading effects. It becomes essential to monitor the rate of flow of information and stop spreading the fake or false messages. The application of suitable graph network modelling and theories would support this research issue and recommend the appropriate model for the social media data.
TopIntroduction
The COVID-19 pandemic increased the number of social media users worldwide in recent days. Data science has a significant role in social media data analytics. Social media becomes the prominent forum for most of the business wherein they adopt digital marketing strategies to attract the new customers and retain the old customers (Darwiesh, 2022). Social media accumulates a large amount of data both in structured or unstructured formats. It becomes the data source for big data analytics like trend analysis, market analysis, sentiment analysis and so on. Data science techniques like data mining, machine learning, deep learning, Artificial Intelligence (AI) and Natural Language Processing (NLP) are widely used in social media analytics. For example, if a company wants to study its consumer behavior to predict its market, then clustering and regression techniques play a crucial role in this forecasting analysis. Social media companies use AI and NLP techniques to analyze users’ profile and their behavior. They develop intelligent algorithms to recognize people in pictures and for content development.
Social media analytics is the process of gathering data from social media platforms like twitter, Facebook, LinkedIn, Reddit and infer information from them. This analytics is highly helpful for businesses in their decision making processes and tracking their performance (Trifiro, 2022). It is not only the count of likes and dislikes, tweets and retweets, and so on. It requires the support of web crawlers which collect data from different channels based on the search queries, then pre-processed, modeled and analyzed for further interpretation.
The success or failure of a new product launch, movie release and others depends on how people react on social media platforms. The bad experience of customer service spread very fast whereas the good messages spread slowly. Organizations have to continuously monitor their brand values and public perception in social media in order to sustain in the market (Chaudhary, 2021). Governments start monitoring the wellbeing of their people and their emotions for the launch of new policies or regulations through social media text and its dashboard (Ismail, 2022). Sometimes, fake news or rumors which spread in internet or social media badly hit the business (Gao, 2022). Machine learning techniques can be used for detection of fake news and the business may take corrective and preventive measures at the right time to protect their customers (Kausar, 2022).
Cyberbullying becomes a serious issue now-a-days in social media platforms. People send abusive or harmful messages or attachments to other users through these digital platforms, which in turn creates psychological impact to the receivers. Social media analytics specific to these type of harassments, comments or tweets along with Natural Language Processing (NLP) algorithms and machine learning techniques like Bayesian analysis help them to trace those senders and it is possible to do further investigations (Perez, 2023; Kausar, 2022).