Kunal Roy

https://github.com/kunalroy**/DataScience.git 310-***-**** adkaor@r.postjobfree.com

Technical Skills

Data Science: Database Management, Data Visualization & Insights, Machine Learning Algorithms, Statistical Modeling & Analysis, Experimental Design & A/B Testing, Model Deployment Programming/Tools: R, Python, SQL, MATLAB, Hadoop, SPSS, Tableau, Google Analytics Libraries: Pandas, NumPy, Matplotlib, SciKit Learn, TensorFlow, PyTorch, ggplot, dplyr, shiny Projects

Classifying Marketing Campaign Responses with Python

- Performed data cleaning by feature engineering, feature scaling, imputing missing values, encoding categorical variables, discretizing variables etc.

- Conducted an exploratory analysis of the data by visualizing and analyzing box plots, histograms, density plots, correlation matrix etc.

- Ran multiple classification models on the data to find the best one by looking at metrics like accuracy, confusion matrix, sensitivity, specificity and ROC curve

- Identified that the model needed a higher specificity for this specific domain problem, also analyzed the AUROC to find out logistic regression was the best model

- Tuned the model performance by adding copies of the underrepresented class to correct class imbalance in the data and improve the sensitivity metric by 38% Predicting the Price of Airbnb Stays with Python

- Created a logistic regression model from scratch using a sigmoid and loss function to predict whether the Airbnb was expensive or not

- Optimized the logistic regression through the gradient descent algorithm by updating weight parameters for 100,000 iterations

- Fitted the data to a linear regression model to predict continuous prices, evaluated the model through cross validation and R-squared error

- Experimented with more powerful models like decision tree/random forest regressor and used the grid search function to automate hyper parameter tuning Other Projects: Artificial Neural Network to Predict Customer Churn with PyTorch, Credit Card Data Cluster Analysis in Python, Anomaly Detection with Isolation Forest in Python, Loan Purchase Classification Pipeline using SciKit Learn, Prediction Model Deployment in Web App using Flask Education

DePaul University, Chicago, IL (Class of 2022)

Master of Science in Data Science w/ concentration in Computational Methods Relevant Coursework: Fundamentals of Data Science, Regression and Statistics, Python Programming, Database Processing, Machine Learning Algorithms, Neural Networks & Deep Learning, Mining Big Data, Time Series Analysis & Forecasting, Computer Vision, Information Retrieval University of California, Los Angeles (Class of 2018) Bachelor of Science in Applied Mathematics

Relevant Coursework: Linear Algebra, Probability Theory, Mathematical Statistics, Calculus, Optimization, Mathematical Modeling, Differential Equations, Mathematical Image Processing, Discrete Mathematics, Graph Theory, Financial Mathematics, Economics, C++ Programming