📂 Pavel Grigoryev - Data Analysis Portfolio

A collection of data analysis projects demonstrating end-to-end analytics capabilities and business impact.

📑 Contents

• 🧑‍💻 About me
• 🛠️ Languages and Tools
• 🎯 Skills
• 😎 Awesome Data Analysis
• 🧩 Building Startup Analytics
• 🌊 Deep Sales Analysis of Olist E-Commerce
• 🌐 WWI Data Pipeline and Dashboard
• 🚀 Frameon - Advanced Analytics for Pandas
• 📜 License

---

🧑‍💻 About me

• I hold a higher technical education.
• I specialize in data analysis with a focus on empowering informed decision-making.
• By extracting insights from complex data sets, I help organizations make data-driven decisions that drive business growth and improvement.

---

🛠️ Languages and Tools

• Programming Languages: Python, SQL (PostgreSQL, MySQL, ClickHouse), NoSQL (MongoDB).
• Data Analysis & Visualization:
  • Libraries: Pandas, NumPy, SciPy, Statsmodels, Pingouin, Plotly, Matplotlib, Seaborn.
  • Tools & Frameworks: Dash, Power BI, Tableau, Redash, DataLens, Superset.
• Big Data & Distributed Computing: Apache Spark, Apache Airflow.
• Machine learning and AI: Scikit-learn, MLlib.
• Time Series Forecasting: Facebook Prophet, Uber Orbit.
• Natural Language Processing: NLTK, SpaCy, TextBlob.
• Web scraping: BeautifulSoup, Selenium, Scrapy.
• DevOps: Linux, Git, Docker.
• IDEs: VS Code, Google Colab, Jupyter Notebook, Zeppelin, PyCharm.

⬆ back to contents

---

🎯 Skills

• Deep data analysis:
  • Preprocessing, cleaning, and identifying patterns using visualization to support decision-making.
• Writing complex SQL queries:
  • Working with nested queries, window functions, CASE and WITH statements for data extraction and analysis.
• Understanding product strategy:
  • Knowledge of product development and improvement principles, including analyzing user needs and formulating recommendations for its growth.
• Product metrics analysis:
  • LTV, RR, CR, ARPU, ARPPU, MAU, DAU, and other key performance indicators.
• Conducting A/B testing:
  • Analyzing results using statistical methods to evaluate the effectiveness of changes.
• Cohort analysis and RFM segmentation:
  • Identifying user behavior patterns to optimize marketing strategies.
• End-to-End Data Pipelines:
  • Building automated ETL processes from databases to dashboards with Airflow orchestration.
• Data visualization and dashboard development:
  • Creating interactive reports in Tableau, Redash, Power BI, and other tools for presenting analytics.
• Web scraping:
  • Experience in extracting data from websites using tools and libraries such as BeautifulSoup, Scrapy, and Selenium for information gathering and data analysis.
• Working with big data:
  • Experience with tools and technologies for processing large volumes of data (e.g., Hadoop, Spark).
• Machine Learning Applications:
  • Capable of building and applying machine learning models for data analysis tasks, including forecasting, classification, and clustering, to uncover deeper insights and enhance decision-making processes.
• Business and Metric Forecasting:
  • Building and interpreting time series forecasts for key business metrics using libraries like Uber Orbit and Facebook Prophet for intuitive, robust forecasting to support strategic planning and goal-setting.
• Working with APIs:
  • Integrating and extracting data from various sources via APIs.
• Process Automation:
  • Automating data workflows and routine tasks using Linux scripting, Apache Airflow and other DevOps tools.

⬆ back to contents

---

😎 Awesome Data Analysis

500+ curated resources for data analysis and data science: tools, libraries, roadmaps, cheatsheets, and interview guides.

View Repository

Key Methods:

• Knowledge Management & Information Architecture:
  • Systematic content curation, resource classification, and learning path development
• Research & Critical Thinking:
  • Technical content evaluation, accuracy validation, and relevance assessment
• Content Strategy & Curation:
  • Quality control implementation, information synthesis, and accessibility optimization

Project Description:

• A curated knowledge hub demonstrating systematic approach to data analysis, reflecting expertise in structuring complex information and evaluating technical content.

Project Goal:

• To create a comprehensive, well-organized resource collection that facilitates learning and professional development in data analysis and data science.

Key Achievements:

• Systematized 500+ resources into logical learning paths and competency areas
• Implemented rigorous quality control by selecting materials based on accuracy and relevance
• Optimized information architecture for quick navigation and knowledge discovery
• Enhanced accessibility through web version development
• Synthesized fragmented knowledge into unified, actionable framework

Business Impact:

• Established trusted reference platform that accelerates learning curve for data professionals and demonstrates expertise in information architecture and knowledge management.

⬆ back to contents

---

🧩 Building Startup Analytics

Building analytics process for startup: infrastructure, dashboards, A/B testing, forecasting, automated reports, and anomaly detection.

View Repository

Stack:

• Data & DB: Python Pandas ClickHouse
• Viz & BI: Superset Yandex DataLens Plotly
• ML & Stats: StatsModels SciPy Pingouin Uber Orbit
• Automation: Apache Airflow Telegram API

Key Methods:

• Data Infrastructure Design:
  • Star schema modeling, ETL pipeline development, and data quality validation
• Product Analytics:
  • Retention analysis, cohort analysis, and engagement metrics tracking
• Business Intelligence:
  • Real-time dashboard design, KPI definition, and self-service reporting implementation
• Statistical Hypothesis Testing:
  • A/A and A/B test analysis, sample size calculation, and statistical power analysis
• Time Series Forecasting:
  • Bayesian structural models, trend/seasonality decomposition, and model validation
• Anomaly Detection:
  • MAD-based outlier detection, alert threshold optimization, and real-time monitoring
• Automation Engineering:
  • DAG orchestration, API integration, and scheduled reporting systems
• Monte Carlo Simulation:
  • Statistical power estimation and sample size determination through simulation

Project Description:

• This project demonstrates the implementation of a complete product analytics system for an early-stage startup that has developed an application merging a messenger with a personalized news feed.
• In this ecosystem, users can browse and interact with posts (views, likes) while simultaneously communicating with each other through direct messages.
• The core challenge was to build the entire analytical infrastructure from scratch to understand user behavior across both features and enable data-driven decision-making.

Project Goal:

• To build complete analytics infrastructure from scratch enabling data-driven product decisions through automated reporting, experimentation, and monitoring.

Key Achievements:

• Built scalable data infrastructure with optimized analytical database in ClickHouse
• Designed interactive dashboards for real-time monitoring of user engagement and retention
• Implemented rigorous A/B testing pipeline with statistical validation framework
• Developed forecasting models for server load prediction and capacity planning
• Created automated reporting system with daily Telegram delivery to stakeholders
• Established real-time anomaly detection for proactive issue resolution

Business Impact:

• Enabled data-driven product decisions and reduced manual reporting overhead through comprehensive analytics ecosystem.

⬆ back to contents

---

🌊 Deep Sales Analysis of Olist E-Commerce

Comprehensive analysis of Brazilian e-commerce data, uncovering key insights and actionable business recommendations.

View Repository

Stack:

• Data Analysis: Python Pandas NumPy
• Visualization: Plotly Tableau
• Statistics & ML: StatsModels SciPy Sklearn Pingouin
• NLP & Text Processing: NLTK TextBlob

Key Methods:

• Exploratory Data Analysis (EDA):
  • Statistical summaries, missing value analysis, and outlier detection
• Data Preprocessing:
  • Feature engineering, missing value handling, and creation of new metrics and dimensions
• Time Series Analysis:
  • Revenue/order trends, seasonality decomposition
• RFM Segmentation:
  • Customer value clustering (Recency, Frequency, Monetary)
• Clustering:
  • sklearn-based customer behavior segmentation
• Geospatial Analysis:
  • Sales heatmaps and delivery performance by region
• NLP Sentiment Analysis:
  • Review text processing with NLTK and TextBlob
• Statistical Testing:
  • correlation analysis and hypothesis testing

Project Description:

• Comprehensive analysis of Brazilian e-commerce platform Olist, identifying growth opportunities and operational improvements through data-driven insights.

Project Goal:

• To perform deep-dive analysis identifying growth opportunities, operational improvements, and customer behavior patterns.

Key Achievements:

• Conducted time-series analysis of sales dynamics, seasonality, and trend decomposition
• Implemented anomaly detection in orders, payments, and delivery times
• Developed customer profiling through RFM segmentation and clustering analysis
• Performed cohort analysis to track customer retention and lifetime value (LTV)
• Processed customer reviews using NLP for sentiment analysis and insights
• Validated business hypotheses through statistical testing
• Delivered strategic recommendations for logistics optimization and sales growth

Business Impact:

• Provided data-backed insights to optimize logistics, enhance customer retention strategies, and drive revenue growth through targeted improvements.

⬆ back to contents

---

🌐 WWI Data Pipeline and Dashboard

End-to-end data pipeline and interactive dashboard for Wide World Importers.

View Repository

Stack:

• Data & Databases: Python SQL PostgreSQL Sqlalchemy DBLink
• Analytics & BI: Yandex DataLens
• Automation: Airflow

Key Methods:

• Database Management:
  • PostgreSQL with OLTP to OLAP transformation
• ETL Pipeline Development:
  • Automated data extraction, transformation, and loading processes
• Data Warehouse Design:
  • Star schema implementation for analytical queries
• SQL Optimization:
  • Complex queries, materialized views, and index optimization
• Business Intelligence:
  • Interactive dashboard development in Yandex DataLens
• Automation:
  • Airflow DAG design for daily data pipeline execution

Project Description:

• End-to-end data pipeline and business intelligence solution for global distributor Wide World Importers.

Project Goal:

• To transform siloed operational data into unified business intelligence platform supporting sales, procurement, and logistics decisions.

Key Achievements:

• Built automated ETL pipeline transforming OLTP data into optimized star schema data mart
• Designed and implemented interactive dashboard for sales, logistics, and customer analytics
• Developed daily automated data updates with Airflow DAG orchestration
• Created comprehensive business intelligence platform with specialized dashboards
• Enabled cross-departmental data-driven decision making

Business Impact:

• Reduced manual reporting time and provided single source of truth for business performance monitoring across departments.

⬆ back to contents

---

🚀 Frameon - Advanced Analytics for Pandas

Frameon extends pandas DataFrame with analysis methods while keeping all original functionality intact.

View Repository

Stack:

• Core Technologies: Python Pandas NumPy
• Statistics & ML: Statsmodels Scikit-learn SciPy Pingouin
• Visualization: Plotly
• NLP & Text: TextBlob
• Documentation: Sphinx

Key Methods:

• Package Development:
  • End-to-end Python package creation and distribution workflow
• Software Engineering:
  • Object-oriented programming and pandas extension development
• Testing & Quality:
  • Automated testing with GitHub Actions and code quality enforcement
• Documentation:
  • Comprehensive documentation generation with Sphinx
• Visualization:
  • Automated chart generation and interactive plotting
• Machine Learning:
  • Feature analysis and model evaluation techniques
• Text Processing:
  • NLP methods for text analysis and sentiment detection

Project Description:

• Powerful pandas extension that enhances DataFrames with production-ready analytics while maintaining native functionality.

Project Goal:

• To create comprehensive analytics toolkit that streamlines exploratory analysis and statistical workflows within pandas ecosystem.

Key Achievements:

• Seamlessly integrates exploratory analysis, statistical testing and visualization into pandas workflows
• Provides instant insights through automated data profiling and quality checks
• Enables cohort analysis with flexible periodization and metric customization
• Offers built-in statistical methods (bootstrap, effect sizes, group comparisons)
• Generates interactive visualizations with single-command access
• Supports both DataFrame-level and column-specific analysis
• Maintains full backward compatibility with native pandas functionality

Business Impact:

• Accelerates data analysis workflows and standardizes analytical methodologies across teams and projects.

⬆ back to contents

---

📜 License

This project is shared under MIT License.