An end-to-end topic-based sentiment analysis engine with deep learning and NLP through top-down way to decipher patient sentiment around issues of interest for venture opportunities sourcing for a data-driven venture capital firm.
Tracking top sector movers based on metrics related to customer consumption to inform potential investment opportunities, identifying top demographic factors of customers correlated to the metric movement, and analyzing structural vs behavioral contribution of representative customer segments to the change.
Analysis on cohort & funnel, segmentation, and the health of revenue model or financial metrics.
Designed online clustering algorithm to identify different situations (i.e. related context groups) based on tweet messages and other features as one of conponents of this DARPA funded project, which is used to inform emergency services and humanitarian aid, providing clues about where to focus rescue efforts.