Latest projects

A project that involves web scraping data from vivino.com, a wine review website. The project collects information about wines and their bottle images and enriches the data with additional details based on the creator's 6-year experience as a sommelier. The project is implemented in Jupyter Notebook.

This project focuses on developing a model to determine the market value of cars for a used car dealership's app. Historical data, including technical specifications, equipment, and prices of vehicles, are available for analysis. The goal is to build an accurate and efficient prediction model while considering prediction quality, prediction speed, and study time.

A machine learning project that aims to build a model for predicting the toxicity of comments. The goal is to quickly respond to toxic comments from users and intercept them in a timely manner. The project uses a dataset of over 150,000 English-language annotated comments, which are categorized into normal comments and toxic comments.

This project uses historical data to predict the number of taxi orders for the next hour at airports. The goal is to attract more drivers during peak periods. The project involves data loading and analysis, model training, and RMSE metric evaluation. The model with the best performance will be selected for deployment.

This project focuses on developing a model to determine the market value of cars for a used car dealership's app. Historical data, including technical specifications, equipment, and prices of vehicles, are available for analysis. The goal is to build an accurate and efficient prediction model while considering prediction quality, prediction speed, and study time.

This project aims to protect the personal data of an insurance company's customers by developing a data transformation method. The method will ensure that the quality of the machine learning models is not affected while preventing unauthorized access to sensitive data. The project involves developing and testing the transformation method.

This project aims to build a prototype machine learning model for an industrial solutions company to predict the gold ore recovery factor. The model will optimize production for profitable outcomes. The project involves data preparation, exploratory analysis, and model building.

This data science project aims to determine the most profitable location for a new well for a production company. Three regions with 10,000 oil fields were analyzed for oil quality and reserve volume. A machine learning model was built, and Bootstrap analysis was used to assess risks and potential gains. The project improved the company's profit by selecting the optimal drilling location.

This project analyzes the Bank Turnover Dataset from Kaggle. Its goal is to predict customer turnover by building and evaluating machine learning models. Models like Decision Tree, Random Forest, Dummy, and Logistic Regression are trained and evaluated using metrics such as accuracy, confusion matrix, recall, precision, F1-score, ROC AUC score, and precision-recall curve. The project enables the bank to manage customer retention and reduce losses proactively.