Hello!!
I am Vishal Gupta, and I am currently a graduate student at UT Austin, doing my Masters’ in Business Analytics graduating in May 2022. I have 4 years of experience as a Data Scientist, primarily in Retail. In the past I have worked with Target (major US retailer) & Zeno Health (Pharmacy retail startup) with focus on Pricing, Supply chain and Operations.
In my professional journey I have gone from contributing as an individual, doing the task to taking initiative and leading the projects where in my latest role I managed a team of two individual. I have also worked along with aspiring Data scientist and mentored them while they worked in my team as Summer Interns.
I love working in a cross-functional team where people with different roles and skillset come together to solve a challenging business problem, having people from different background brings in different perspective and accelerate innovation. I believe not every problem can be solved with Data but there are a lot of them that can be and I try to learn new tools and techniques that can enable me in doing the same.
Projects
Magellan Health Capstone
Working with Magellan Health as Data Science Capstone Intern I, along with a team of 5 other individuals, are helping them build member level claim prediction model from scratch that can estimate future claim for each member and the total pharmacy cost for a prospective client.
Brand Engagement Analysis on Social Media - Nike vs Adidas
All the Instagram post of Nike and Adidas were analyzed - Image (video thumbnail) and Caption, to provide insight into the significantly higher reach of Nike with 177M followers vs Adidas with only 26M followers. Important attributes of a post are identified by obtaining labels from image (Google Vision API) and text mining on the caption. The dimensionality of the data is reduced using Topic Modeling to perform a supervised learning with Likes & Comments as predictor variable. Regression analysis is performed to understand the topics and attributes of post that can be predictive of number of likes.
Crowdsourced Recommendation Engine for Beers
User feedback is essential for characterizing a product and target the product to prospective consumers. Beer recommendation engine utilizes customer reviews for different kinds of beer, and a specific brand is suggested to a user that combines the three attributes most important to him/her/them. The recommendation engine uses a two-pronged approach to arrive at the closest match using Sentiment score (vaderSentiment analyzer) and Similarity score (cosine similarity & spaCy). Review similarity score is obtained by identifying the presence of beer attributes from a manually created list.
Competitive analysis of Luxury Car makers from discussion forums
Text mining of user generated content coupled with cluster analysis can help brand identify competitors as perceived by the customers. Comments & post from a car discussion forum are scrapped and analyzed to identify the group of car makers that people perceive to be similar. Lift analysis is done to identify brands associated frequently and Multi-dimensional scaling plots are created to visualize proximity and cluster of car makers.
Hospital Length of Stay Prediction in ICU
Resource management is essential in health care operations and COVID-19 exposed the necessity for better estimation to manage resource allocation and capacity. Length of stay for a patient is bucketed into 11 different buckets and different classification models, Logistic Regression, Decision Tree, Random Forest and Naïve Bayes, are leveraged to predict the bucketed patient length of stay at the ICU at the time of admission.
Code | Presentation | Kaggle Problem Statement
Optimization Projects
The repo contains projects code and report for Optimization part 1 and part 2 offered in MSBA at UT Austin.
Work Experience
Zeno Health
Lead Data Scientist April 2021 - July 2021
Senior Data Scientist February 2020 - March 2021
- Demand Forecasting and replenishment engine
- Supplier recommendation for purchase orders
Target
Senior Data Scientist June 2019 - January 2020
Data Scientist July 2017 - May 2019
- Fleet sizing for general merchandise and food distribution center
- Price recommendation for competitive assortment using history and competitor data
- Digital orders (Target.com) profile generation for fulfillment center load estimation
Patent
SYSTEM AND METHOD FOR MANAGING TRANSPORTATION VESSELS