George Kacoyanis – Project Portfolio

India T-Shirt Sales Data Exploration in PowerBI

Using PowerBI to analyze the revenue of T-Shirts in India. Created several measurements using DAX to create Month over Month Time Series growth between 2017-2018. Used a Year slicer, a pie chart, filter chart, Scorecards. Used Power Querying and Dax to crete a date table and a relationship of 1 to many between the dates in the date table and order dates.
View Project

Machine Learning Data Exploration in Excel

Exploratory data analysis for the Boston Housing Area to find the drivers in Median Value of Owner-Occupied Homes, using Frontline Data Solver in Excel to conduct these methods: Principle Componenet Analysis, Correlation Analysis, Regression analysis (Multi-Variate Linear and Logisitc), Regression and Classification Trees, Random Trees, Neural Networks, K-Nearest Neighbors Means Clustering, and boosting/bagging classification and regression ensembles.
View Project

ARIMA Forecasting using SAS

Forecasting transportation patterns and variable temperatures using SAS to apply ARIMA models with seasonal differencing, log transformation, and residual diagnostics to forecast monthly temperature and transportation time series data, selecting models based on AIC/SBC and validating with autocorrelation checks and confidence intervals.
View Project

Image Denoising using Bayesian Analysis

A Bayesian image denoising project using Markov Random Fields, Gibbs Sampling, and Metropolis-Hastings. Replacing different sigma values from the proposed ones, from 0.01 to 1, to find the best sigmas for denoising the image.
View Project

PowerBI and Python Data Visualizations

Using Python and PowerBI to create serval data visualizations using multiple datasets. Implemented PowerBI functions such as DAX and Power Querying.
View Project

Python Machine Learning Projects

The projects consist of methods using matplotlib and numpy libraries to create conditionals, loops and functions, plots, Monte Carlo simulation of random processes and for business applications, polynomial regression, k-means clustering, And Artificial Neural Networks.
View Project