Artyom Ashigov

Iris Species Predictor (ML in Production Project)

Built and deployed a simple ML app with Streamlit. Users can download a sample CSV, upload data, and get class probabilities plus the predicted Iris species.

Analysis of Stolen Vehicles in New Zealand

Performed data analysis in R programming language on stolen vehicle data from New Zealand. The project examined theft patterns by weekday, vehicle type, region, color, vehicle age, and population density, with visualizations and insights to support conclusions.

CI/CD Pipeline with GitHub Actions and AWS EC2

Built a CI/CD project using GitHub Actions, a self-hosted runner on AWS EC2, Docker, FastAPI, and pytest. The pipeline automatically builds the app, runs tests, and redeploys the latest version to the server after each code push.

Sentiment Analysis of the Top Armenian Banks

Developed a sentiment analysis project in Python using web scraping, AWS Translate, AWS Comprehend, and AWS S3 to analyze news articles from leading Armenian banks. The project explored public sentiment across institutions through multilingual text processing, cloud-based analysis, and visualizations.

Serverless Market Data Pipeline on AWS

Built an end-to-end serverless AWS data pipeline that ingests stock market data from an external API using AWS Lambda, stores raw JSON in Amazon S3, transforms nested data with AWS Glue, exposes curated Parquet data through Amazon Athena, and visualizes the final dataset in Power BI. The pipeline is orchestrated with AWS Step Functions and scheduled automatically with Amazon EventBridge.

Business Aviation CO₂ Emissions Analysis

Developed a Python-based analysis to estimate CO₂ emissions for business aviation flights. Implemented a multi-source fuel burn estimation approach using a 4-level fallback cascade (internal benchmarks, industry reports, public datasets, and aircraft-level approximations) to maximize data coverage.

Handled incomplete flight data by calculating distances using the Haversine formula and deriving missing flight times based on aircraft cruise speeds. The project focused on building a reliable and scalable methodology for emissions estimation in real-world, imperfect datasets.

Based on proprietary data, no public dataset available.

UK Bank Customer Demographics & Financial Analysis

Developed an interactive Tableau dashboard with story-driven navigation to analyze customer demographics and financial behavior across UK regions. The project explored regional distribution, balance segmentation, age and gender patterns, and job classification to uncover insights into customer profiles. Implemented dynamic filters and storytelling views to highlight key trends such as regional demographic differences and customer segmentation.

Bike Usage & Weather Impact Analysis

Built an interactive Tableau dashboard to analyze bike usage patterns, station activity, rider demographics, and weather impact. The project explored monthly trip trends, top stations by trip volume, geographic trip distribution, age distribution, and customer versus subscriber behavior to identify usage patterns and support operational insights.

Operational & Pricing Analytics in Delivery Business

Completed a data analysis case study involving SQL querying and Tableau visualization to evaluate delivery performance, pricing behavior, and operational efficiency. Built a suite of interactive dashboards analyzing order volume, delivery time, pricing distribution, and store-level performance. Identified key insights such as peak demand hours, delivery time variability across stores, and pricing behavior across regions, supporting data-driven decision making.