1. PIXIUAI
[This project is a part of my tenure at Pixiu as a founding data scientist trainee]
Problem
Only 27% of Indian adults and 24% of women meet the minimum level of financial literacy as defined by the Reserve Bank of India.
Financial illiteracy is an increasingly large problem in the United States, a fact especially true for minority groups in lower socioeconomic classes. According to a recent survey conducted by Standard & Poor's, only 57% of U.S. adults are financially literate.
Solution
Built a fintech community that keeps one updated with finance news, where one can learn finance and the basics of investment through courses and being a part of the community
Implementation
Worked on finding different Twitter handles that provide finance-related information
Research on different news sources that post about investment and crypto
Scrapped data and annotated into 3 broad categories for twitter data
-- News
--Advertisement
--Greetings
Evaluation
A machine learning model was trained for a classification task to display real-time relevant news
ML model evaluation is dependent on a fixed range for accuracy, precision, and recall for deployment.
Results
I implemented it for finding relevant news and courses for finance in the first launch.
This has later been implemented for crypto
Uniqueness
Built a platform that combines finance-related courses/ information/news with an online community where one can discuss their queries regarding investment
Being a founding member, the demands from the work expanded to technical, social, and community building skills