This project aims to analyze historical data from the "StreamChick" online store, specializing in global computer game sales. By examining factors such as user ratings, genres, and platforms, the goal is to uncover patterns that contribute to a game's success. Through exploratory data analysis and hypothesis testing, we seek to identify potentially profitable products and formulate strategies for future advertising campaigns, ultimately enhancing decision-making processes within the gaming industry.
This project revolves around assisting the mobile operator "Megaline" in recommending appropriate tariffs, "Smart" or "Ultra", to its customers based on their behavior data. Utilizing preprocessed customer data, the aim is to develop a classification model achieving a minimum accuracy of 0.75 on the test set. The dataset is divided into training, validation, and test subsets for model evaluation, where different models with varied hyperparameters are explored to optimize accuracy. As an additional task, model adequacy is examined to ensure reliability, acknowledging the potential complexity of the dataset.
This project aims to forecast taxi orders at airports for the upcoming hour to help a company, "Fair Taxi," efficiently allocate resources during peak demand periods. Historical taxi order data is analyzed and various machine learning models are trained and tested using different hyperparameters. The goal is to achieve a Root Mean Square Error (RMSE) metric value on the test dataset of no more than 48. Through this process, the project seeks to optimize driver allocation and improve service efficiency based on predictive analytics.