loandangnt Goto Github PK

I’m a Data Insight Analyst based in Ho Chi Minh City with a passion for discovering customer/product insights through data analytics. Leveraging my background in economics and business management, I help data-driven companies tell actionable stories. I love chatting about women leadership and analytics. In my spare time, I do embroidery and coloring.

Topic modeling customers' text review for clothing items. 23,486 reviews were preprocessed using NLTK and RegEx. The analysis resovled the questions of what customers like and dislike about each clothing category, as well as identifying what problems need prioritized for a particular clothing item that was dislike (negatively reviewed, lower than 3 in a scale of 5). The unsupervised Latent Dirichlet Analysis topic modeling technique achieved 77% accuracy based on human eyes verification. The model result was a list of the most alarming issues that needs improvements for each negative text review.

Black Friday Sales Prediction predict based on past purchase, customer demographics and product details. Techniques used: XGBoost, Linear Regression, K-means Clustering. Ranked 470 in public LB.