Social listening tool helped client with finding top influencers, sentimental analysis etc, with the help of twitter data using ML Algoriithms
Sentiment analysis tool that uses twitter data and applies SVM and Naive Bayes ML algorithms to help client see the top influencers, competitor intelligence, effects of various campaigns, desired product features and time to time sentiment of their brand.
- The Client is a leading telecom company providing telecommunications and IT services to corporate clients
- The Client wants to leverage social media to improve their business
- The Client wants to know what different topics and trends are discussed online with respect to their business and their competitor’s business
- The twitter handles of the client and the competitors, for a given date range, is stored in an S3 bucket using an AWS API Gateway
- The tweets are analysed and sentiments for tweets are created using an ensemble of SVM and Naive Bayes Machine Learning algorithm. This completes the Social Listening Process
- Next, topic modelling is done using an LDA algorithm that creates a mapping between the different topics discussed in the tweets and the keywords used to influence the same
Visualization – Social Listening
Visualization -Topic Modelling
- Sentiment Analysis helped the client to identify the mood prevailing in the minds of the customers about their product and services. Around 10,000 tweets were analysed, wherein 70% of them had a neutral outlook
- Competitor Analysis gave insights about their competitors and the mindset of the customers about the competitor products and services and it showed similar trend of major sentiment being neutral
- Top Influencers, Competitor Intelligence, Effects of Various Campaigns, Important Issues, Desired Product Features, Time to Time Sentiment were identified through this analysis
- Topic Modelling helped the client to quickly understand the topics discussed in the tweets and address the issues if there were any
If you want to leverage social media to improve their business, let’s connect!