This repository contains reports for some of the personal projects I worked on. The purpose is to demonstrate my approach for solving complex problems using real-world data. Code chunks for every step are included. Data can be found in the related github repositories along with source code and R markdown files used to make the reports.
Summary: The goal of this project was to build a prototype for cocktail recommendation app. The recommendation is made using cocktails’ ingredients and a list of user rated cocktails. Additionally, the prototype is capable of unsupervised clustering of cocktails based on their similarity to each other.
Link: Cocktail Reccomendation App
Summary: The goal of this project was to asses financial viability of applied machine learning in sports betting domain. For prototyping and training of machine learning algorithms historical data for soccer games played in LaLiga division (since 1993) was used. Betting odds of the most recent games from 03/15/17 till 05/21/17 were used to estimate Return on Investment (ROI). Best models demonstrated 8 to 13% ROI using features engineered by calculating simple statistics which represent time-weighted performance of each team and include: average number of goals scored, average number of goals scored playing at home and away, average number of games won/lost and some of their derivatives.
Link: LaLiga Sports Betting
Summary: Can brain activity be transformed into music? EEG from Neurosky Mindwave Mobile was converted into notes and played using Overtone/SuperCollider audio engine.
Link: EEG Brainwave Music