In a world that is optimising for quantity of relationships through “People You May Know", I wanted to build a product that is focusing on the quality of relationships. I believe that a meaningful relationship happens because of shared passion, not because of the school you went to or how many mutual friends you have. The lucky thing about our human nature is that we are wired to make connections between dots and make introductions. For example, if both person A and person B want to empower young girls to learn science. Once our brain correctly retrieved those information, we can make introductions. The challenge is that we might miss such opportunities by only relying on our memories, especially for Super Connectors who know thousands of people from different circles and are the keys of connecting 2 circles together or building a new community. (Go to this link for a graphical explanation of how super connectors contribute to high quality human connections: http://tinyurl.com/anatomy-of-social-network). After a month of intensive focused learning, fast prototyping and data mining, I built a machine learning model that extracts topics from messy conversational text data to help my client, a Super Connector, to discover new introductions he could make in minutes.
Incorporated NLP techniques, topic modeling, feature extractions, and network graphs.
Tools used are: Python, genism, sk-learn / scikit-learn, HTML, Latent Semantic Analysis, Latent Dirichlet Allocation, and Non-Negative Matrix Factorization, NetworkX, NLTK, Spacy, Pandas, Seaborn and Matplotlib.
Skills demonstrated: product management, data analytics, Web Scraping, story-telling, user research, data mining, data munging, ETL, data visualisation, machine learning, and statistical analysis.
Target Audience / Stakeholder: SuperConnectors