A chatbot that takes care of people during a lockdown.

A sentiment analysis model that responds to emotion.

Lais Carvalho

APIs Natural Language Processing Web Servers and MicroFWs (Flask/Tornado/Nginx/...) python

ICare is a chatbot that helps people overcome loneliness. We started this project because we wanted to help people who suffer from loneliness during the COVID-19 social distancing times. The app allows people to give ‘voice’ to their emotions - especially the negative ones, without relying on their friends or (negatively) influencing on their moods.

The chatbot originally uses Facebook Messenger’s built-in NLP to extract entities from people’s ‘emotive messages’ and classify them according to the system’s intent. This pre-built model turned out to be quite limited, triggering responses only for sad or happy sentences. Although very suitable for e-commerce and business in general, this option was not suitable for our aim.

We decided then to use Flask to build an application that will use a modelling tool for sentiment discovery and emotion classification to analyse the user’s input sentences and suggest an emotion such as anger, fear, joy, etc. that would output an appropriate response to the user.

Type: Talk (30 mins); Python level: Intermediate; Domain level: Intermediate

Lais Carvalho

Python Ireland

I’m a 3rd year IT student in CCT Colege Dublin that fell in love with python after struggling with Wolfran for 2 years while studying calculus for engineering. I worked for the Brazilian government designing water damns for least-privileged communities in the Northeast of the country before deciding to move to Ireland. With 6 years of direct experience with customer service, I am currently training as a Remote Conference Operator with projects throughout the tech community such as Pyjamas and Python Ireland, aiming to contribute more significantly to the community.