R, SQL, Python, Machine Learning, Data Visualization
Resume:
MITUL SHAH
EDUCATION
Master of Science in Data Science (3.80/4.0) Michigan Technological University, Houghton, MI Dec 2017 Relevant Coursework: Data Mining, Multivariate Statistics, Advanced Statistical Analysis, Predictive Modelling, Data Visualization, Machine Learning (Stanford University), Analytics Edge (MIT), Database Management Bachelor of Electrical Engineering (3.80/4.0) Gujarat Technological University, Gujarat, India May 2013 TECHNICAL SKILLS
• Programming Languages: R, Python, SQL, MATLAB, Apache Spark
• Tools: Revolution R Enterprise for Big Data, Weka, Tableau, Adobe Illustrator
• Theoretical Knowledge: Machine Learning, Descriptive and Inferential Statistics, Probability, Linear Algebra WORK EXPERIENCE
Data Scientist Learning & Decision Lab, Houghton, MI Sept 2017-current
• Collected pupil data for different participants using Tobii X3-120 by taking care of the experimental method.
• Performed data cleaning and analyzed the pupil sizes for participants in different conditions. Research Assistant Harold Meese Centre, Houghton, MI Sept 2017-current
• Working on different applications of measuring divergence and correspondence between the paths with Dr. Shane Mueller
(the author of the library pathmapping in R) to show how this library can be used in different domains. Research Assistant Harold Meese Centre, Houghton, MI March 2017-May 2017
• Predicted mask (an image that appears immediately before and after the stimulus to disrupt processing) with more than 90% accuracy by building different classification models using stratified cross-validation in a psychology experiment. ACADEMIC PROJECTS
Optimizing Conversion Rate Fall 2017
• Inspected data quality issues and visualized different variables to explore their impact on the conversion rate.
• Built a Random Forest model to predict conversion rate and improved the accuracy from 97% (baseline model) to 98.5%.
• Derived insights from the model to help the product and the marketing team in improving the conversion rate. Engagement Test Summer 2017
• Tested the performance of a new feature added on the site by looking at the performance of different user segments.
• Chose not to add the feature on the site by concluding it to be a novelty effect. Marketing Email Campaign Summer 2017
• Built a model to send emails in a new optimized way and estimated an increase in the click-through rate of about 50%.
• Suggested an approach to test the new model and discovered an interesting pattern in one of the segments of users. Funnel Analysis Spring 2017
• Merged 4 datasets to find a drop in the conversion rate time series for the mobile and the desktop users.
• Provided some insights on what the product team should focus on to improve the conversion rate. INDEPENDENT PROJECTS
Spanish Translation A/B Test Summer 2017
• Applied t-test to find out a drop of 10% on conversion rate in the new version of the site.
• Discovered the bias in the experiment by utilizing the power of a decision tree as a descriptive tool. Ads Analysis Spring 2017
• Identified the 5 best performing ad groups by choosing the right metric to improve engagement.
• Clustered ad groups into 3 groups: the ones whose average cost per click was going up, was going down and was flat. Fraud Detection Spring 2017
• Merged 2 datasets about the user information and their ip addresses to identify the user country.
• Predicted fraud with about 99.6% accuracy using Random Forest to use the model from a product perspective. 251 Ariba Drive,
Sunnyvale, CA
******@***.***
linkedin/mitulpshah
github/mitulpshah
Contact this candidate