Employee Network Analysis

Data Analytics

This project was created for DataCamp competition on network analysis. The aim of this study is to create an employee network using message data and use this network to gain insight on relationships among employess and departments. Below are a few important points.
The project includes several features to analyze the employee network:
Network visualization: A non-directed weighted graph is produced using the NetworkX library. The graph nodes contain information on different aspects of the network, such as the departments or ages of the employees.
Data Analysis: Relationships between messages and age, department and date are explored using graph data.
Centrality metrics: Several centrality metrics are calculated including degree centrality, betweenness centrality. These metrics are to identify the most important employees in the network.

Chess Set Classification

Machine Learning

The Chess Piece Image Classification is a deep learning project aimed at classifying images of different chess pieces using convolutional neural networks (CNNs). Dataset consists of over 7,000 images of six different types of chess pieces, including the king, queen, rook, bishop, knight, and pawn.
The project begins with data preprocessing, including resizing and reshaping the images, as well as applying various image augmentation techniques to increase the size of the training dataset. The deep learning network used in the project includes a simple CNN with four convolutional layers and a 2 linear layers. The network was trained and evaluated using various performance metrics, including accuracy, precision, recall, and F1 score.

Movie Recommendation Using NMF

Machine Learning

This project provides a recommendation system for movies based on Non-negative Matrix Factorization (NMF). The processing steps are:

*Text preprocessing (Using spacy, nltk)
*Vectorization (Using Tf-Idf Vectorization)
*Non-Negative Matrix Multiplication
*Calculation of movie similarity from final feature matrix

Credit Card Data Clustering

Data Analytics

This project aims to perform clustering analysis on credit card data to identify patterns and group similar customers together.
The dataset contains information about credit card holders, including their credit limits, balances, payment status, and demographic data.
Hierarchical Clustering, K-Means, Gaussian Mixture and DBScan are applied

A/B Testing for Website Redesign

Data Analytics

This project provides an analysis of an A/B test performed on a website redesign for a Datacamp Competition.
Python pandas, scipy and statsmodels packages are used to perform a hypothesis test on conversion statistics
The analysis includes 2 main parts.
Statistical analysis: The script performs several statistical tests to compare the performance of the old and new versions of the website.
Effect size calculation: The script also calculates effect sizes for the statistical tests. This can be used to quantify the magnitude of the differences between the old and new versions of the website.

BTK Telecom Data Extraction

Data Pipeline and Processing

This project is aimed at extracting information from BTK Turkey yearly telecom statistics and market datasets and making a visual analysis.
The yearly data was extracted using Python libraries, preprocessed and cleaned to remove duplicates and missing values. Then all data was merged on a year an location basis
PowerBI was used for data analysis and visualization on a city and product level. The project's findings show patterns in the telecommunication data, including trends in consumer behavior and preferences.

NBA Player Data in PostgreSQL

Data Pipeline and Processing

The EDA Istanbul City Theaters project is a data analysis project aimed at exploring the performances of the Istanbul City Theaters, a network of public theaters in Istanbul, Turkey. The project includes a dataset of over 27,000 performances spanning from 2002 to 2019, including the date, location, and category of the performance, as well as the number of tickets sold. The goal of the project is to gain insights into the trends and patterns of the Istanbul City Theaters' performances over time, as well as to identify the most popular genres and venues. Various exploratory data analysis techniques are applied, such as visualizations and statistical tests, to uncover interesting patterns and relationships within the data. The project also includes machine learning models for predicting play category.

IBB Istanbul Weather and Traffic Analysis

Data Analytics

This project is an analysis of the Istanbul Weather and Traffic Data based on datasets extracted Istanbul Metropolitan Municipality data portal.
Analysis consists of 2 main steps:
Exploratory Data Analysis: The relationship among variables are explored, such as average temperature, number of vehicles. Various statistical methods and data visualization techniques are used to gain insights into the data.
Time Series Forecasting for Weather: ARIMA Model is used for time series forecasing to predict the daily average temperature

Real-Time Order Processing Pipeline with Kafka & PostgreSQL

Data Pipeline and Processing

This project implements a real-time data pipeline for processing new orders using Apache Kafka, PostgreSQL, Docker, and Python.
The system periodically polls an API to fetch new orders and publishes them to a Kafka topic. A Kafka consumer then reads these messages, extracts timestamps and prices, and appends them to a CSV file. The project is containerized using Docker to ensure seamless communication between components within a dedicated network.
The architecture is designed for scalability and fault tolerance, making it suitable for real-time order tracking and analytics. This setup can be extended to integrate with Big Data frameworks, ETL processes, or business intelligence tools for further analysis.

TicTacToe-Reinforcement-Learning

Machine Learning

The TicTacToe Reinforcement Learning project is a machine learning project aimed at teaching an agent how to play Tic Tac Toe using reinforcement learning.
Python player and game modules are used to train an agent. The agent is trained against a random player, allowing it to learn the optimal moves and strategies through trial and error.
The project does not have a user interface but game module that allows users to play against the trained agent. The training process can also be observed in Jupyter notebook.

Exploratory Data Analysis - Istanbul City Theater Plays

Data Analytics

This project is a data analysis study aimed at exploring the performances of the Istanbul City Theaters, a network of public theaters in Istanbul, Turkey. The dataset contains information about over 27,000 performances spanning from 2002 to 2019, including the date, location, and category of the performance, as well as the number of tickets sold.
The goal of the project is to gain insights into the trends and patterns of the Istanbul City Theaters' performances over time, as well as to identify the most popular genres and venues. Various exploratory data analysis techniques are applied, such as visualizations and statistical tests, to uncover interesting patterns and relationships within the data. The project also includes machine learning models for predicting play category.

End-to-End Crypto Data Pipeline with Azure and Databricks

Data Pipeline and Processing

Cloud News Data Pipeline - DBT&Airflow

Data Pipeline and Processing