Addressing a significant industry challenge, this project effectively predicts customer churn in the telecom sector using machine learning. The top-performing model, an XGBoost classifier, achieved an impressive 92% Recall Score in predicting positive cases of "churn" with other impressive performances in other evaluation metrics used too. Insights from this initiative empower the industry to implement strategic retention measures, mitigating churn and boosting customer satisfaction in a sector grappling with multimillion-dollar losses each month due to customer attrition.

This project allows businesses to customize their offerings, products and services to different customer segments, instead of utilizing a generic approach. Through techniques like unsupervised learning, I gain a better understanding of customers, facilitating adjustments to products based on specific needs, behaviors, and concerns of various customer types. For instance, instead of allocating resources to market a new product to every customer in the company's database, I can analyze which customer segment is most likely to make a purchase and then concentrate marketing efforts on that particular segment.

This project harnesses the power of data and AI to forecast and analyze employee attribution trends. Gradient Boosting Algorithms including XGBoost, CatBoost, LightGBM and an ensemble of these models were used to develop robust predictive models that provide insights into employee retention, identify patterns, and contribute to strategic talent management.

I wrangled, analyzed and visualized the tweet archive of Twitter user @dogrates also known as WeRateDogs. WeRateDogs is a Twitter account that rates people's dogs with a humorous comment about the dog. I wrangled WeRateDogs Twitter data to create interesting and trustworthy analysis and visualizations.
The Twitter archive is great, but it only contains very basic tweet information. I then got additional information by querying Twitter's API to gather more data.

In this project, I analyze the Prosper loan dataset using Python to create insightful visualizations. The process involves thorough exploration, data cleaning, and preparation to extract meaningful insights into borrower behavior and loan performance. The objective is to comprehend patterns and trends in the Prosper loan data, covering loan origination rates, borrower credit scores, and critical metrics. These insights are invaluable for informed lending decisions, risk management strategies, and financial analysis. The project serves as a testament to the effectiveness of data visualization in simplifying complex data for a broader audience's understanding.

This project was part of my Data Analyst Nanodegree Programme at Udacity. The soccer database, sourced from Kaggle, is ideal for data analysis and machine learning, containing comprehensive data on matches, players, and teams across European countries from 2008 to 2016. It involved thorough data analysis using SQL and Python, including data wrangling, querying with SQL, and further analyses using Python. Through this project, I gained valuable insights into soccer data patterns and trends, encompassing player performance, team rankings, and other vital metrics. These insights are instrumental in guiding coaching decisions, player recruitment, and other critical aspects of soccer management.

As a football fan, it was always clear that I would apply data science in some way to the sport. This marks my first venture into web scraping. In this project, I'm scraping data from the La Liga football league website using Python's Beautiful Soup library. The scraped data encompasses player statistics, team rankings, and various other crucial metrics tied to the league. The main aim of this project is to collect data for further analysis and to gain insights into player and team performance throughout the season. It's a great showcase of the potential of web scraping for efficiently gathering large volumes of data, presenting a valuable tool for football enthusiasts and analysts alike.