Projects

In this project I use a python library called selenium to automate a repetitive task.

I found myself stuck with a tedious task of describing each subject from my university academic history. So, I put my Python skills to use along with ChatGPT, and created a program that automates the process. The program logs into the university website, navigates to the page displaying the offered subjects for the production engineering course, and clicks on each of the 130 subject rows. It then checks for a description and extracts the required information from each subject.

Extracting relevant data manually from numerous emails and entering it into spreadsheets can be time-consuming, error-prone, and inefficient. This is where Python, a versatile programming language, becomes incredibly valuable.

 In this project, I applied data science techniques to analyze a dataset containing information about different beauty products and customer reviews. The objective was to perform a sentiment analysis on different categories of shampoo to gain valuable insights and provide data-driven recommendations.

The project:

1. Dataset: The dataset was obtained at Kaggle from Nykaa, an indian #beauty and #fashion e-commerce. The dataset included product details, customer reviews, ratings, and more. It covered a wide range of beauty products from various brands. #pandas library was used to process the data and extract only information related to shampoo products.

2. Categorization: To enhance the analysis, I categorized products based on specific attributes, such as argan-based products or strawberry-scented shampoos. This categorization allowed for a deeper understanding of consumer preferences and trends.

3. Sentiment Analysis: One of the key techniques employed was sentiment analysis. By using natural language processing tools like #TextBlob library, I evaluated the sentiment polarity of customer comments. This helped understand the overall sentiment associated with different categories of shampoo.

4. Trimmed Mean Analysis: Additionally, I conducted a trimmed mean analysis to calculate sentiment scores while excluding extreme values. This approach provided a more robust and representative measure of sentiment for each category.

5. Insights and Recommendations: I used #powerbi to visualize the data. By analyzing the data and sentiment scores, I uncovered valuable insights, such as the brand performance, customer satisfaction levels, and the most positively reviewed products. These insights can be used by beauty companies to make informed decisions regarding product development, marketing strategies, and customer engagement.