| Summary: | Social content is available everywhere on the Internet today. Performing text analytics provides meaning to the unstructured data gathered from the social content. It is opportune to tap into online content given how the social media is thriving worldwide currently, generating more data than before. To aid businesses in market research using automation, two main ideas were explored in this project. Specifically focusing on data regarding travel and insurance, topic modelling by Latent Dirichlet Allocation (LDA) and Sentiment Analysis were implemented on forum posts data and Twitter data respectively. The results derived from LDA picked out a few topics that were general features that travel insurance encompassed and could be extended to other areas of concern, while sentiments towards and against similar brands in the insurance industry were studied to provide a visualisation of how members of the public view these competing brands. These analytics provided insight to the collaborating company that would otherwise require more manpower to collate and analyse. More work could be done in terms of tailoring the models to the insurance industry. These tested methods could be adopted for similar use cases accordingly.
|
|---|