Welcome to my portfolio repository! This is a collection of the data science and machine learning projects I've worked on, showcasing my skills in data analysis, predictive modeling, and machine learning. Each project demonstrates different aspects of my capabilities, from data preprocessing to model deployment.
- Credit Card Fraud Detection
- Titanic Survival Prediction
- Iris Flower Classification
- Sales Prediction Using Python
- About Me
- Contact
In this project, I developed a machine learning model to detect fraudulent credit card transactions. The goal was to create a system that accurately identifies fraud while minimizing false positives, which is crucial in real-world financial applications.
The dataset consisted of anonymized credit card transactions, with features such as transaction amount, time, and more. I employed techniques like SMOTE to address the class imbalance and cleaned the data to ensure accuracy.
- Logistic Regression
- Random Forest Classifier
- SMOTE (Synthetic Minority Over-sampling Technique)
- Cross-validation for model evaluation
The final model achieved an accuracy of 99% and a precision of 98%, effectively reducing the number of false positives while maintaining high accuracy in detecting fraudulent transactions.
This project used the famous Titanic dataset to predict the survival of passengers based on features like age, gender, class, and more. The focus was on applying various classification algorithms and feature engineering techniques to improve prediction accuracy.
The dataset included information on Titanic passengers, such as age, gender, fare, and whether they survived. I performed extensive feature engineering, including handling missing values and creating new features like family size.
- Logistic Regression
- Support Vector Machines (SVM)
- Decision Trees
- Feature engineering and selection
The model was able to achieve a high level of accuracy, with a final score of 82%, by effectively utilizing both numerical and categorical features in predicting passenger survival.
This project involved building a model to classify iris flowers into three species: Setosa, Versicolor, and Virginica. It was a straightforward classification problem that helped reinforce my understanding of fundamental machine learning concepts.
The Iris dataset contains 150 samples of iris flowers with four features: sepal length, sepal width, petal length, and petal width. No significant preprocessing was required, making it a clean and ideal dataset for classification tasks.
- K-Nearest Neighbors (KNN)
- Decision Trees
- Support Vector Machines (SVM)
- Cross-validation for model tuning
The model achieved an accuracy of 96%, demonstrating the effectiveness of simple yet powerful classification algorithms on well-structured datasets.
This project aimed to forecast future sales based on historical data. The objective was to build a predictive model that can assist businesses in making informed decisions about inventory and marketing strategies.
The dataset included past sales data with features such as date, store, product, and sales volume. I performed time series analysis and used data preprocessing techniques like normalization and feature extraction.
- Linear Regression
- ARIMA (AutoRegressive Integrated Moving Average)
- Time Series Analysis
- Feature engineering and selection
The model successfully predicted future sales trends with a high degree of accuracy, providing valuable insights that can help in decision-making processes for inventory management and marketing.
I am Nisar Ahmed Siddiqui, a data enthusiast with a Bachelor's degree in Artificial Intelligence and Data Science. My passion for uncovering patterns and insights from data has led me to work on a variety of projects, each of which has helped me build and refine my skills in data science, machine learning, and analytics.
I am currently pursuing an internship where I continue to learn and apply these skills in real-world scenarios. I am always eager to take on new challenges and collaborate with others to drive impactful results.
Feel free to connect with me: