💻 Personal Projects

💼 My Data Science, Machine Learning & Software Projects

🩺 Diabetes EDA & Visualization

Explored the Pima Indians Diabetes Dataset focusing on:

• Glucose level, BMI, Age patterns
  
  

• Correlation analysis, distribution plots
  
  

• Tools: Python (Pandas, Matplotlib, Seaborn)
  Outcome: Identified key health factors influencing diabetes probability using visual analytics.
  
  

🚗 Car Ownership Prediction

Developed a predictive ML model to understand car ownership behavior.

• Features: Income, Age, Location
  
  

• Techniques: Regression Models, Feature Engineering, Scaling
  
  

• Tools: Scikit-learn, NumPy
  Outcome: Predicted car ownership likelihood based on demographic and socioeconomic data.
  
  

🧹 Data Cleaning & Business Statistics Project

Performed complete data preprocessing and statistical analysis.

• Removed duplicates, handled missing values
  
  

• Applied summary statistics, outlier detection
  
  

• Tools: Pandas, Matplotlib, Excel
  Outcome: Improved dataset quality and generated practical business insights.
  
  

🚕 Uber Data Analysis – Dhaka City

Analyzed synthetic Uber ride data:

• Peak hour detection
  
  

• Zone-wise demand analysis
  
  

• Visuals: Heatmaps, Bar charts, Pivot tables
  Key Insights: Highest demand between 8–10 AM & 6–8 PM; Busiest areas — Gulshan, Dhanmondi, Mirpur.
  
  

🤖 ML Classification – SVM (RBF Kernel)

Implemented Support Vector Machine Model for classification.

• Dataset: Public binary-class dataset
  
  

• Applied RBF kernel, scaling, parameter tuning
  
  

• Tools: Scikit-learn, Matplotlib
  Outcome: Achieved stronger accuracy after kernel & hyperparameter optimization.
  
  

📈 Regression Models & Visualization

Worked on various regression approaches:

• Linear Regression
  
  

• Polynomial Regression
  
  

• Visualizations: Scatter plots, Trend lines
  Outcome: Showed how scaling & model complexity affect prediction accuracy.
  
  

🧬 Imbalanced Data Handling

Focused on solving class imbalance issues.

• Oversampling (SMOTE)
  
  

• Undersampling
  
  

• Model performance comparison
  Outcome: Improved classification results with balanced datasets.
  
  

📝 Spam Detection Project

Developed a machine learning model to classify email/SMS as spam or not.

• Text cleaning, TF-IDF vectorization
  
  

• Algorithms: Naive Bayes / SVM
  Outcome: Achieved high accuracy with optimized text preprocessing.
  
  

🧮 Superstore Sales Analysis & Forecasting

Analyzed sales trends and created forecasting models.

• Monthly, yearly pattern analysis
  
  

• Top/Bottom profitable products
  
  

• Used ARIMA / Time-series forecasting techniques
   

📊 Business Statistics – Data Cleaning Project

• Applied business metrics
  
  

• Conducted descriptive & inferential statistics
  Outcome: Produced clean datasets ready for decision-making.
  
  

Github Account :   github.com/mahin020202