London, UK
Data Analyst &
Data Science
MSc Graduate.
Machine learning · NLP · GIS
Emil Niyazov - Data Analyst and Data Science MSc Graduate with Distinction
Making sense of data.
Making something useful.
I focus on turning complex data into clear, decision-ready insights. My work spans real-estate analytics, NLP for document intelligence, and spatial dashboards. I enjoy projects that combine data engineering, visualization, and automation to create practical, measurable impact.
Work history.
- Machine Learning & NLP: Built anomaly detection and classification models on property listings and images, improving the accuracy of real estate analytics from 80% to 98%.
- Pipeline Engineering: Engineered automated pipelines for web scraping, transformation, and SQL storage, ensuring clean and reliable analytical datasets, and reduced manual review time by 5 hours/week.
- Dashboards & GIS Tools: Designed Tableau dashboards for HR and financial teams and developed an internal GIS application to enhance spatial analysis capabilities.
- Reporting & Growth: Delivered concise real estate market reports with key trends and forecasts, contributing to strategic planning while progressing from Intern to Junior and then Data Analyst.
- Automated Data Collection: Built Python scrapers (Scrapy, Pydoll, Playwright) pulling from 40+ international news, government and multilateral sources, with per-site parsers and handling for dynamic pages and rate limits.
- Data Cleaning & Structuring: Developed a pandas processing layer for de-duplication, boilerplate and paywall stripping, language detection and country/entity tagging, normalising output into a queryable PostgreSQL corpus.
- AI Analysis & Automation: Designed the prompt architecture and output schema for an LLM stage (OpenAI API, locally hosted Ollama models) producing source-cited assessments of economic, political and geopolitical conditions across 6 countries; scheduled in Docker, replacing ~8 hours of manual research per market.
- Strategic Research & Speech Development: Conducted in-depth research to shape impactful themes and talking points for the Secretary General's speeches at high-level international meetings.
- Event & Stakeholder Coordination: Coordinated the X and XI Global Baku Forums, managing diverse stakeholders and ensuring seamless planning and execution.
- High-Level Engagements: Participated in delegations meeting with global leaders, presidents, ministers, and senior officials, contributing to strategic activities during High-Level Meetings.
Selected work(10)
LibrAIs - AI Video Intelligence for YouTube
A web app that takes a YouTube link and automatically transcribes it, generates subtitles, summarises the content, extracts key insights, and even identifies the University of Sheffield building where the video was recorded - helping users quickly find the most relevant content.
Baku Primary Market Residential Report 2022-2025
Interactive Tableau story analyzing the primary residential market in Baku with trends, price dynamics, and key metrics for 2022-2025.
SEC 10-K Annual Reports Analyzer
Web app that parses and analyzes SEC 10-K reports to extract key financial and risk information using NLP.
Data-Driven WebGIS Application
2D/3D maps, property visualization, measurements, attribute tables, layer filtering, Excel/PDF exports, image-rich pop-ups. Login: test / test123.
Azerbaijani Carpet Dataset on Hugging Face
Curated dataset of Azerbaijani carpet images for computer vision tasks: classification and generative modeling.
Real Estate Analysis Using Spark RDD
Large-scale real-estate data analysis using Apache Spark RDDs for price modeling and feature engineering.
Hour of Code
Volunteer activity teaching programming basics as part of the Hour of Code initiative.
Factors Related with Depression
Study of adults 50+ in England. Chi-square tests and logistic regression show depression is strongly linked to general health, life satisfaction, loneliness, and internet use patterns.
Audio Characteristic Analysis of Popular Songs
Billboard Year-End Top 100 (2000-2023) analysis via Spotify features. PCA + k-means clustering revealed four song archetypes - no single "formula" for a hit.
Predicting House Prices - Data Mining & AI
Compared multiple linear regression and random forest on 1,300 Ames housing records using KNIME, model-specific preprocessing, and 10-fold cross-validation. Held-out test R² reached 0.863 for linear regression and 0.862 for random forest on log-transformed prices, showing comparable performance despite the difference in model complexity.
Education & Certifications
MSc in Data Science
Graduated with Distinction
University of Sheffield
Sheffield, United Kingdom · Sep 2025 - Sep 2026
Dissertation
Comparative Study of Transformer-E and TCN-AE Architectures for Time Series Anomaly Detection
Focus: Transformer and temporal convolutional network (TCN) analyses.
BS in Computer Science
ADA University
Baku, Azerbaijan · Sep 2021 - May 2025
Certifications
I International Internet Olympiad · May 2022
XI Global Baku Forum · NGIC
Team LibrAIs · The University of Sheffield · 2026
Technical Stack
Books I Learn From
Hands-On Machine Learning with Scikit-Learn, Keras & TensorFlow
Aurélien Géron · Practical ML patterns and workflows.
Pattern Recognition and Machine Learning
Christopher Bishop · Core theory for probabilistic ML.
Python for Data Analysis
Wes McKinney · Data wrangling with pandas.
Deep Learning
Goodfellow, Bengio & Courville · Foundational deep learning reference.
Storytelling with Data
Cole Nussbaumer Knaflic · Communicating insights clearly.
Let’s make
something matter.
Have a project or opportunity in mind? Let’s talk.