Zoya Marium Bhaju
Data Analyst
3+ projects
Python
SQL
Data Wrangling & Cleaning: 6/10
Exploratory Analysis & Visualization: 6/10
Business Insight & Impact: 5/10
Active 5 hours ago
+91 (7595) 935885
Invite to interview
Message
Download CVCV
Overview
Technical skills
Timeline
Roles
Overview
A data engineer-level practitioner focused on end-to-end analytics and ETL pipelines with hands-on experience in SQL, Pandas and BI workflows. The strongest proven skill is practical data ingestion and cleaning as shown by the SQL null-checks and prod_Churn table creation plus the CSV-merging and parsing pipeline in the sales analysis notebook. Public artifacts do not show predictive modeling, automated testing, containerized or productionized pipelines, or environment/version management.
Technical skills
Python• Middle • 3 projects
SQL• Middle • 3 projects
Databases
Databases
MySQL• 3 projects
AI/ML
NumPy
Pandas
Jupyter Notebook
DevOps
GitHub
Analytics
ETL/ELT
Seaborn
Matplotlib• 3 projects
Power BI• 3 projects
Timeline
University of Engineering and Management (UEM Kolkata University)
Bachelor's Degree •
Computer Science and Engineering (Artificial Intelligence and Machine Learning)
Analyzed 186,000+ sales transactions using Python, Pandas and Matplotlib to identify revenue trends and customer purchasing patterns. Performed data cleaning, feature engineering, and exploratory data analysis by creating Month, Hour, City, and Sales features from transactional data. Identified peak sales periods, optimal advertisement timings, top-performing cities, and best-selling products through statistical analysis and visualizations. Conducted market basket analysis to determine frequently purchased product combinations and provide cross-selling recommendations.
Python
NumPy
Pandas
Matplotlib
Developed an interactive Customer Behavior Analysis Dashboard using Power BI to analyze customer purchasing patterns, subscription trends, revenue distribution, and demographic insights. Performed data cleaning in Python, EDA in MySQL and visualization in Power BI to generate business insights and support decision-making. Created KPI cards, slicers, bar charts, and donut charts to visualize metrics like total customers, average purchase amount, average review rating and category-wise sales and revenue. Leveraged DAX measures and dynamic filtering to uncover business insights related to customer segments, product categories, and sales performance for data-driven decision-making
Python
Power BI
MySQL
GitHub
SQLAlchemy
Developed an end-to-end Customer Churn Analysis Dashboard in Power BI to analyze customer behavior, churn trends, revenue contribution, and service usage patterns. Performed ETL operations including data extraction, cleaning, transformation, and loading of customer data into MySQL. Executed SQL queries to analyze demographics, contracts, payment methods, and churn categories. Performed transformation in Power Query. Created DAX measures for KPI cards such as Total Customers, Total Churn, Churn Rate, and New Joiners. Built interactive Power BI dashboards with filters, tooltips, and visual insights to support customer retention strategies and data-driven business decision-making.
Power BI
MySQL
ETL/ELT
Data Science Intern
•
Junior
1Stop.ai
•
Internship
Python
NLP
Statistics and Machine Learning Toolbox
Middle Data Scientist
Confidence: Medium Data Engineer
A data engineer-level practitioner focused on end-to-end analytics and ETL pipelines with hands-on experience in SQL, Pandas and BI workflows. The strongest proven skill is practical data ingestion and cleaning as shown by the SQL null-checks and prod_Churn table creation plus the CSV-merging and parsing pipeline in the sales analysis notebook. Public artifacts do not show predictive modeling, automated testing, containerized or productionized pipelines, or environment/version management.
Statistical Rigor
2/10
Correct use of statistics
Limited statistical rigor: descriptive summaries and visual checks appear but no hypothesis testing, uncertainty quantification, or formal inference.
Evidence
Sales-Data-Analysis-using-Python/SalesAnalysis.ipynb: uses df.describe() and visual inspection without statistical tests
customer_behavior_analysis/Customer_behavior_analysis.ipynb: median imputation and descriptive checks but no significance testing
Data Wrangling & Cleaning
6/10
Preparing and cleaning data
Solid data wrangling and cleaning practices for single-node pipelines: null checks and IFNULL handling in SQL, merging many CSVs, parsing/coercing dates, type coercion and group-based imputation in notebooks.
Evidence
Customer_Churn_Analysis/Customer_Churn_Query.sql: extensive null-count checks and CREATE TABLE prod_Churn with IFNULL defaults
Sales-Data-Analysis-using-Python/SalesAnalysis.ipynb: concatenates 12 monthly CSVs, drops NaNs, coerces dates with errors='coerce' and converts numeric types
customer_behavior_analysis/Customer_behavior_analysis.ipynb: group-median imputation, renaming to snake_case, qcut age grouping and mapping purchase frequencies
Exploratory Analysis & Visualization
6/10
Exploring and visualizing data
Exploratory analysis produces clear visualizations and business-oriented interpretations; charts are used to answer specific questions and include concise written takeaways.
Evidence
Sales-Data-Analysis-using-Python/SalesAnalysis.ipynb: monthly_sales bar chart, city sales bar chart, hourly orders line chart with written conclusions about ad timing
Sales-Data-Analysis-using-Python/SalesAnalysis.ipynb: product-pair analysis using combinations and Counter to find frequently co-purchased items
Predictive Modeling
1/10
Building models that predict
No predictive modeling or model evaluation present; work focuses on EDA, SQL analytics and ETL rather than ML.
Evidence
customer_behavior_analysis/Customer_behavior_analysis.ipynb: ends after EDA and MySQL upload, no modeling or cross-validation steps
Sales-Data-Analysis-using-Python/SalesAnalysis.ipynb: analysis and visualization only, no predictive model code
Business Insight & Impact
5/10
Turning analysis into business value
Business-focused analysis: revenue calculations, segmentations and actionable recommendations (ad timing, best months, product bundle insights) are present and tied to metrics.
Evidence
Sales-Data-Analysis-using-Python/SalesAnalysis.ipynb: revenue by month and city, explicit recommendation about best hours to run ads
Customer_Churn_Analysis/Customer_Churn_Query.sql: revenue by Customer_Status and churn category computations
Reproducibility & Notebook Hygiene
2/10
Clean, repeatable analysis
Reproducibility is limited: notebooks contain hard-coded local paths, inline pip installs and plaintext DB credentials; no environment pins, requirements file or data/versioning seen.
Evidence
customer_behavior_analysis/Customer_behavior_analysis.ipynb: hard-coded Windows file paths, inline pip install and plaintext MySQL credentials in create_engine call
Sales-Data-Analysis-using-Python/SalesAnalysis.ipynb: local path variable for CSV folder and no environment specification
Expertise
Analytics• All Tiers
Industries
Commerce• Middle
Data & Analytics• Middle
Telecommunications• Middle
Technologies
Databases
Jupyter Notebook
Pandas
ETL/ELT
Power BI• mentioned only
Python• mentioned only
Recommendations
- Develop and maintain single-node ETL pipelines and SQL-based data transformations for analytics workloads using MySQL and Pandas.
- Build business dashboards and KPIs in Power BI while providing the cleaned, aggregated data via SQL tables or simple API endpoints.
- Implement repeatable data ingestion jobs that read monthly CSVs, validate and coerce types, and write to a staging schema with clear null-handling rules.
- Hardening: remove plaintext credentials from notebooks, add requirements/environment files, and convert exploratory notebooks into parameterized scripts for production scheduling.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories:
Junior AI/ML Engineer
Confidence: Medium Data-centric
Data-focused analyst (Middle) with practical experience building end-to-end data cleaning and ETL flows for retail and churn use cases. The strongest proven skill is data pipeline and feature engineering demonstrated by notebooks that merge monthly CSVs, clean and typecast fields, derive features (age_group, purchase_frequency_days), and write cleaned tables to MySQL. There is no evidence of formal model training, experiment tracking, deployment, or reproducible production pipelines in the public human-authored code.
Model Architecture & Training
How well models are designed and trained
Not evidenced in public code
Data Pipeline & Feature Engineering
4/10
How data is prepared for models
Evidence of end-to-end data preparation and feature engineering: ingestion and merging of multiple CSVs, null handling and type conversions, mapping and derived features, column renaming, and writing cleaned tables to MySQL.
Evidence
customer_behavior_analysis/Customer_behavior_analysis.ipynb: df['review_rating'] imputed using groupby.transform(lambda x: x.fillna(x.median()))
customer_behavior_analysis/Customer_behavior_analysis.ipynb: frequency_mapping dict and df['purchase_frequency_days'] creation
Sales-Data-Analysis-using-Python/SalesAnalysis.ipynb: merging 12 monthly CSV files into all_data.csv and cleaning Order Date, numeric conversions, and City extraction
Experimentation & Evaluation
How results are measured and tested
Not evidenced in public code
MLOps & Deployment
How models are shipped to production
Not evidenced in public code
Computational Efficiency
How efficiently computing resources are used
Not evidenced in public code
Research Depth & Innovation
Depth of research and new ideas
Not evidenced in public code
Industries
Commerce• Middle
Telecommunications• Middle
Technologies
NumPy
GitHub
Python• mentioned only
Recommendations
- Build and productionize repeatable ETL pipelines (e.g., using Airflow or Dagster) that replace notebook-based ad hoc scripts to increase reproducibility.
- Implement feature-store patterns or structured feature engineering scripts so downstream model training can be automated and tracked.
- Develop lightweight model training and evaluation notebooks with clear train/validation splits and basic experiment tracking (MLflow or W&B) if moving toward predictive modeling.
- Replace hardcoded local paths and credentials with configuration files or environment variables and add data/input validation to improve security and portability.
Repositories
The developer's experience in this domain has been verified based on AI analysis of the following repositories: