Welcome to the Skinlytics repository dedicated to data projects related to the main Skinlytics project. This repository focuses on the development of a comprehensive skincare analysis system, encompassing various components and stages. The main objective is to create an advanced web-based dashboard using Dash Plotly that provides interactive visualizations and data-driven insights for skincare products.
Live version of the dashboard can be accessed here: http://anaritasantos.pythonanywhere.com/
Interactive table displaying the targeted product list, including brand, product name, full ingredient list, price, quantity, and price per ml. Additional columns, such as review sentiment and main product benefits based on the ingredient list, will be progressively added throughout the project.
Access and preprocess the data, ensuring cleanliness and consistency in the ingredient lists. Implement a classification system to categorize ingredients based on their role and function. Generate spider radar visualizations to represent the ingredient profiles of each product.
Develop a sophisticated recommender engine based on a similarity matrix, taking into account common active ingredients among products. Enhance the recommender engine with more review-generated variables and adjustable parameters to improve accuracy and personalization.
Create a comprehensive Product Analysis Dashboard using Dash, incorporating the spider radar chart, recommender, and Dash DataTable. This interactive dashboard will serve as the central hub for skincare analysis, providing users with an intuitive and insightful interface.
- Collect and transform data from various sources, including customer reviews, expert opinions, and social media platforms.
- Implement advanced sentiment analysis and topic modeling techniques to extract valuable insights from the reviews.
- Integrate data exploration visualizations into the web dashboard to provide users with a holistic understanding of product reviews and feedback.
- Throughout the project, a strong emphasis will be placed on scalability, performance, and usability to handle a significant volume of skincare products and user interactions.
Listing of sources I used to help me with domain knowledge, data cleaning and code for visualizations and the recommender system.
- Ingredient List Research & Classification: https://incidecoder.com/ https://cosmetics.specialchem.com/ https://www.hautschutzengel.de/
- Product Recommender System: https://www.kaggle.com/code/eward96/skincare-recommendation-engine
- Dash web application programming: Dash Tutorial videos on Youtube (CharmingData channel: https://www.youtube.com/@CharmingData) Dash bootstrap Cheatsheet: https://dashcheatsheet.pythonanywhere.com/
- Other Resources:Dash Documentation, Medium Articles, Chat-GTP