Projects

Cloud based ETL for Ad Analytics

An end-to-end ETL pipeline to analyze YouTube ad trends, helping companies optimize placement and improve campaign results.

Web-app: Real-time Emotion Recognition

A web app containing a computer vision model that recognizes your emotions in real-time through a web camera. The model was crafted through transfer learning applied to ResNet-50 and assembling my own dataset. The trade-off between decent accuracy (81.94%) and the required computational load was reached by only training half of the original network structure. The app utilizes WebSocket to handle asynchronous interaction between the front and the back.

Real-Time KPI Dashboard

A real-time dashboard made for the University of Michigan Ecological Center and integrated with Google Sheets to automate KPI tracking.

Cross-Market Insurance Purchase Prediction

Project goals: 1. to find patterns of customer cross-market purchases based on EDA and Visualizations; 2. Prepare the dataset for model building; 3. To find an ML algorithm that provides better binary predictive results out of five: Logistic Regression, Decision Tree, Random Forest, Gradient Boosting, and XGBoost.

E-Commerce Platform Customer Behavior Analysis

This project was to capture the insightful details of user behavior over one month of October on an E-Commerce Marketplace through EDA and ML model building.

PostgreSQL 13 | Python | Numpy | Pandas

Real Estate Database Building and Analytics

This project's objectives were as follows: 1. To design a conceptual model (ER diagram) of a potential Database that can be made up of several smaller datasets taken from the real estate company Zillow; 2. To translate a conceptual model into an SQL schema by using DDL Statements; 3. To create a PostgreSQL Database and populate it by using DML Statements; 4. To write SQL queries to analyze the tendencies of the real estate market; 5. To analyze the performance of SQL queries;

Tableau | Tableau Prep Builder

Coffee Shop Chain BI Analytics

I helped a small batch coffee roasting company conduct a business analysis based on their data about customers, employees, stores, and products in order to operate the business more effectively through data-driven business decisions.

Tableau | Tableau Prep Builder

Kohl's Storage Optimization

Here we can see two goals: 1. To analyze business requirements and constraints; 2. To conduct the exploratory data analysis to formulate recommendations on the items allocation optimization in the storage.

SAS Enterprise Miner | Tableau | Tableau Prep Builder

Allstate Insurance Policy Purchase Prediction

Objectives: 1. To find out what characteristics define a potential customer mostly; 2. To highlight the features of the policy that correlate with a customer's decision to purchase a policy; 3. To build a binary predicting model to forecast the policy purchasing based on the given features;

Excel

Steel Dynamics, Inc. Financial Analysis

I statistically examined Steel Dynamics financial and business performance compared to their competitors based on the next financial metrics: Contribution Margin Ratio, Operating Leverage Effect, Debt to Equity Ratio, and Financial Leverage.

Kirill Nartov