JEROME JOSHUA
Data Scientist at Circle K- Claim this Profile
Click to upgrade to our gold package
for the full feature experience.
-
English Full professional proficiency
-
Tamil Native or bilingual proficiency
-
Hindi Professional working proficiency
Topline Score
Bio
Credentials
-
CORE JAVA
Oracle -
DATA SCIENCE IN R
Jigsaw Academy
Experience
-
Circle K
-
Retail
-
700 & Above Employee
-
Data Scientist
-
Jun 2021 - Present
• Led a project to build better test control mapping for complex tobacco category, utilizing store similarity algorithms to identify and group similar stores, thereby increasing the accuracy of the measurement process. • Built a demand forecasting project utilizing statistical modeling techniques for demand-driven pricing & and optimized inventory management. • Enhanced clustering-based “localized pricing” scalable solution in Python and Spark to recommend item prices based on estimated price elasticities, which lifted the average margin of 7% in Texas division stores. • Segmented stores using clustering, to implement better strategies to increase profit margins and store traffic. • Conducted ad hoc analyses using SQL, and Python to measure test versus control stores’ impact of price changes on various business metrics like units sold, margin and revenue, and basket metrics and to detect outlier stores. • Built and maintained automatic ETL pipeline across billions of transactional data to measure weekly test vs. control store's performance of localized pricing with lift metrics to support leadership’s decision-making. Show less
-
-
-
LendingTree
-
United States
-
Financial Services
-
700 & Above Employee
-
Data Scientist
-
Aug 2017 - Jul 2019
Worked for an online lending exchange company that compares various products from banks and publishes financial blogs that help consumers to choose the right decision. • Developed a customer targeting model to identify potential customers to purchase a credit card specific to Subprime and Near prime customer segments and optimize the marketing cost of direct mail campaigns by 16%. This model was developed in Python, relies on Logistic regression, was trained on 10 million credit report data with 22 features finalized, and achieved an accuracy score of 81%. • Developed and deployed a classification model to predict consumers who are likely to purchase a student loan by their web behavior from Google analytics data and helped the forecasting team provide expected revenue daily and strategize marketing campaigns immediately , reducing the time period from 30-45 days of realizing actual revenue. This model was developed in Python, relies on random forests, trained on google analytics data with 7 features, and achieved an accuracy of 77%. • Provided insights and mortgage trends by performing data analysis on 150 million households in R language, to provide interesting content for blog pages which increased the organic web traffic of the mortgage section of the business site by 35%. • Implemented and maintained moving average forecasting model for weekly, monthly & quarterly in R language, predicting the page views & clicks of the business site for each financial product through Google analytics data and help the analytics team to strategize each product's marketing and increase revenue. • Conducted exploratory data analysis and infer insights on USA’s public and financial data to fuel our business site blog pages resulting in the increase of business site organic traffic and thereby increase revenue saving online marketing cost. Show less
-
-
Education
-
2019 - 2021
University of Michigan
Master's degree, Data Science -
2013 - 2017
Anna University
Bachelor of Technology, Information Technology
Community
