ASHISH GUPTA https://www.linkedin.com/in/gupta-ashish-01

Hartford, CT 860-***-**** • ac0gcl@r.postjobfree.com 2+ years of experience in statistical analysis, modeling, data visualization and BI. Command in R, Python, Tableau. Expertise in Insurance and Marketing domains. Exposure to various Industry relevant data science/analytics projects EDUCATION

Master of Science in Business Analytics & Project Management, GPA 4.0/4.0 April 2017 University of Connecticut, School of Business Hartford, CT Bachelor of Technology in Mechanical Engineering, Delhi Technological University, Delhi – India May 2014 TECHNICAL SKILLS, CERTIFICATIONS AND COURSEWORK

Data Science & Visualization Tools: Python [Base, GraphLab create, SFrame], R Programming, R shiny, Tableau, SQL, PL/SQL, SAS [Base, EG Miner JMP], MS Excel [Analytical Solver], Google Analytics, Hadoop MapReduce, Pig, Hive, Scala. Statistical Skills and Machine Learning: Hypothesis testing [T-test, ANOVA, Chi-Sq], Probability distributions, A/B Testing, Market basket Analysis, Principle Component Analysis, Clustering and Factor Analysis, Linear and Quadratic Discriminant Analysis, Regression [Linear, Non-linear, K-NN], Classification [Logistic, tree-based, SVM], Regularization [LASSO, RIDGE], Ensemble modeling [Bagging, Boosting, Stacking], Neural Network, Evaluation metrics, Natural Language Processing, Text analytics, Time Series Forecasting, ARIMA, ARIMAX. Relevant Certifications and Coursework: Machine learning Foundations, Regression, Classification (university of Washington - Coursera), The Analytics Edge (MIT – edX), Predictive Modeling, Business Intelligence and Data Mining, Big Data Hadoop, Data Management and Process Modeling, Business Decision Modeling (Linear Programming), R Programing, Data Visualization – Tableau/MS Excel.

PROJECTS IN DATA ANALYTICS/SCIENCE

• Click-Through Rate Prediction, Avazu: R, Python – Analyzed online advertising with data of over 40 million rows. Applied feature engineering (Hashing, Impact coding) to capture interactions. Built models like logistic regression with regularization, random forest with bagging and boosting techniques, etc. Achieved logarithmic loss of 0.397.

• Denoising Dirty Scanned Documents, Image Processing and Machine Learning, Adobe Systems: R – Incorporated several Image Processing techniques like Edge Detection, Image Morphology, Median Filter, and Adaptive Thresholding. Implemented XGBoost and Gradient Boost models on engineered features to clean-up noisy text images.

• Created a Rating Plan based on Historical Auto Claim, Kangaroo Auto Insurance: R, SAS JMP – Built boosting models based on zero-inflated Poisson distribution and Tweedie distribution for creating rating plan. Improved Gini by 14%.

• Panel Data Analysis, Rossmann Store: R, SAS – Predicted sales for 1115 stores using different panel data estimation techniques via PLM package in R. Compared different panel data models, and finally forecasted sales using fixed estimation technique with an accuracy of 0.934.

• Reverse Engineered Power BI Dashboard to Recreate in Tableau, Visual Analytics: Tableau – Addressed the rising drug abuse problems and deaths in the state of Connecticut through powerful visualizations. RELEVANT EXPERIENCE

UCONN GRADUATE CONSULTANT LIMRA, Hartford, CT AUG 2016 - Present

• Retirement Planning Research - Building predictive models for clients in financial services using R, SAS EM, SAS JMP.

• Identified business factors impacting surrender of variable annuity products with different rider benefits (e.g. VAGLB).

• Analyzing churn behavior of different age group customers via interactive visualization in R-shiny (ggplot) and Tableau.

• Providing insights and explanations of customer surrender behavior, marketing strategies and fee pricing recommendations to reduce the frequency of new policy surrenders.

• Status– Achieved recall of 87% and AUC of 0.79 in rare class classification problem. 20% improvement over status quo. Data Science Intern Progressive Insights, Boston, MA May 2016 – July 2016

• Identified factors influencing “Crop Yield” via several predictive models to address certain issues in agriculture domain.

• Implemented parametric and non-parametric ML algorithms with modified cost function as per business requirement.

• Built a shiny web application to interactively visualize the crop yield with identified major influencing factors. ENTREPRENEUR Essentials I-CLEAN Pvt. LTD., New Delhi, India June 2014 – October 2015 As the Founder and Director, established the startup with a total of 10 employees. Given the small team, I wore multiple hats. As a team lead, I standardized all processes from start to end, developed business strategies to capture the market, and set the ball rolling through effective leadership and communication. As a BI Analyst and Data Modeler,

• Extracted, merged and analyzed data from a wide variety of sources. Conducted market research, ROI/IRR analysis, Break-Even analysis, A/B testing and hypothesis testing.

• Developed several web-based applications and Tableau dashboards to track KPI’s relating to sales revenue, customer retention and work-flow efficiency.

OTHER ACHIEVEMENTS

• Publication – “Experimental Investigation on Hydrodynamic Journal Bearing using SAE10W30 Oil”, International Journal of Advanced Research and Innovation, paper no. 131, ISBN ***-**-****-328-0.

• Test Score – GMAT - 710.