Sephora Data Analysis For Select ELC Brands
The skills we demoed here can be learned through taking Data Science with Machine Learning bootcamp with NYC Data Science Academy.
Background and Introduction
Estee Lauder Companies (ELC) is a high-end beauty company that was founded in 1946 and became a publicly traded corporation in 1995. Currently, data shows ELC has 30 beauty brands in its portfolio.
Since going public, ELC has heavily focused on growing their brand portfolio through several acquisitions. However, ELC is still focused on maintaining and growing legacy brands (i.e., brands that were part of the portfolio prior to going public), as they are vital to the corporation.
For the purposes of this project, I made the assumption that ELC leadership would like to gain insight into candid customer feedback for ELC’s legacy brands in order to track and maintain brand integrity. The brands in focus for this analysis include:
- Estee Lauder
- La Mer
- Bobbi Brown
To gather candid customer feedback, I performed web scraping of Sephora.com using Selenium and Python.
Data Gathering and Analysis
From Sephora.com, I gathered the following information for each of the products aligned to the five legacy brands listed in the last section:
- Reviews Complete
In total, information aligned to 517 products across the five legacy brands was gathered.
The main analysis was focused on the distribution of product rating vs. product price for each brand. Overall, it was clear that each brand had a distinct price range and generally consistent review ratings for their products. La Mer had the highest average product rating and smallest distribution range for product rating. However, La Mer also had a significantly higher average price point per product. In contrast, Origins had the lowest average product rating and lowest average price point per product. This shows that higher prices are loosely correlated to higher product ratings.
To see additional information regarding my analysis and web scraping code, please visit my Github project.