Imamverdi Badalov JUnior Data Scientist

As a dedicated Data Scientist, I am deeply passionate about Data Science and Artificial Intelligence, with a particular interest in understanding their mathematical underpinnings and logical structures to enhance my expertise and impact.

Fraud Detection

• Data Preprocessing: Performed correlation analysis to identify and remove redundant features. Addressed class imbalance using SMOTE and undersampling techniques.
• Model development: Implemented and evaluated the following models: Gradient Boosting Classifier, Random Forest Classifier, Logistic Regression, Support Vector Classifier, and K-Nearest Neighbors Assessed performance using metrics like accuracy, confusion matrix, and classification report.
• Best Model Selection: Identified the best-performing model based on accuracy and conducted detailed analysis.

Churn Analysis

• Data Preprocessing: Cleaned and prepared the dataset for training. Created a heatmap to visualize correlations or features.
• Model Creation: Built a neural network model with three hidden layers, using ReLU activations and a sigmoid output.
• Model Compilation: Compiled the model with Adam optimizer and binary cross-entropy loss, aiming to predict churn probability.
• Cross-Validation: Applied 5-fold cross-validation and evaluated the model using ROC AUC scores, achieving an average of 80 recall.

Netflix Movie Analysis with MYSQL

• Analyzed order delivery times by payment type, using JOIN to combine orders and payments data.
• Calculated maximum payment revenue by customer state, excluding unavailable or canceled orders.
• Identified orders with multiple payments using LAG and LEAD window functions.
• Merged customer, order, and payment data, and converted payment values to integers.
• Categorized payment types into numerical values for easier analysis.
• Conducted movie name word frequency analysis, breaking down titles and counting word occurrences.
• Standardized movie titles with 'Vol 1' and 'Vol. 2' to 'Part 1' and 'Part 2'.
• Compared movie release years with previous years using LAG for trend analysis, SQL, joins, window functions, recursive queries, CASE statements, data aggregation, cleaning

Power BI Dashboard Project – Local Car Market Analysis

• Collected raw vehicle listings data from turbo.az and cleaned/structured it using Python (Pandas, NumPy).
• Built an analytical dataset for market insights by removing inconsistencies, standardizing price/mileage fields, and organizing location-based attributes.
• Designed interactive Power BI dashboards to help entrepreneurs and small automotive businesses understand local car prices, regional trends, demand indicators, and model-level comparisons.
• Provided data-driven insights for decision-making using visualizations such as price distributions, geographic breakdowns, and time-based market shifts.

• Datalent - Data science
• Coursera - Supervised Machine Learning: Regression and Classification
• Coursera- Advanced learning algorithms
• Technest - Artifical Intelligence

• English, Advanced
• Azerbaijani, Native
• Russian, intermediate