Live app here!
Helps you understand how users are responding to a comment on Hacker News using natural language processing and machine learning. This client is built using React/Redux and relies on a node server
The main idea with this app was to pass the comments in a hacker news thread through a machine leanring model to calculate "sentiment" - a measure of how positive or negatively people responded to the original post. In the early stages of development, I relied on Google's cloud natural language service, but recently deployed a custom model to GCP using TensorflowServing on Kubernetes.