Sentissimo - Sentiment Analysis


Many groups are underrepresented or discriminated against within companies. Because of that, the aim of this project was to create a system that could determine the sentiments of employees at a granular level based on things like ethnicity, gender, and marital status. This data can be used by companies to identify potential areas of improvement, or by potential employees to determine which company is the best fit for them.

Natural Language Processing

We leveraged Amazon's Comprehend system to analyze the sentiments in email messages and mocked a company database that would match these emails to the user's demographic data. We then aggregated this data to determine various averages relating to different demographic groups and presented them in an interface that served to rank companies in terms of Overall Sentiments and Diversity. By using various filters, a more granular view can also be used to view sentiments of different groups over time at the company.

Share this project: