Roushan Kumar

959-***-**** ac0s0t@r.postjobfree.com www.linkedin.com/in/Roushan Kumar

PROFESSIONAL SUMMARY: An analytics professional and a hands-on technologist who enjoys finding stories in data and using statistical & quantitative techniques for solving business problems in banking, supply chain, sales and marketing.

EDUCATION:

University of Connecticut School of Business Hartford, CT

Master of Business Analytics and Project Management, STEM Major, GPA 3.83/4.0 Expected Jul 2017

Concentration: Predictive Modeling, Machine Learning, Data Visualization

Visvesvaraya Technological University Belgaum, India

Bachelor of Engineering in Information Science Jul 2012

Concentration: Database Management, Operations Research, Object-Oriented Programming

Certifications: SAS certified statistical business analyst; Google Analytics

SKILLS AND EXPERTISE

●Programming Languages: SQL, R, SAS, Python, Hadoop, Pig, Hive, C#

●Tools: MS Excel(Advanced), Tableau, MS Visio, MS PowerPoint, MS Visual Studio, SPSS

●Statistics and Machine Learning: Hypothesis Testing, Clustering, PCA, Linear/Logistic Regression, Decision Trees, Bagging, Boosting, Random Forests, Neural Networks, Time Series Forecasting, Bayesian Statistics, SVM, Data Mining, Text Mining, Data Visualization, Cost-benefit Analysis, Sentiment Analysis, Business Intelligence, Recommender systems, A/B Testing

EXPERIENCE

Sterlite Tech – Data Analyst (Data Science) R Tableau SQL Excel 2016

Employed statistical and quantitative techniques such as lasso regression and linear optimization to reduce the scrap and thus increase the yield in the fiber manufacturing process by 6%.

Merged the data from Salesforce(CRM) and SAP HANA(ERP) and built linear regression model to predict the receipt of Purchase Order for each Request for Quotation. The model helped us achieve higher sales funnel conversion rate.

Used recency, frequency and monetary (RFM) indices to perform customer-segmentation using k-means and k-medoids clustering. The approach helped us identify high-churn customers and thereby reduce attrition.

Performed data transformation on the OFC exports data & used two sample t-test and bootstrapping to identify pricing as the prime reason for brand switching of customers. Suggested marketing and negotiation as ways to reduce churn by 5%.

Tata Consultancy Services - Business Analyst (Banking, Pharma) R Tableau SQL C# 2012-2015

Performed data aggregation and clustering and created a time series forecasting model using ARIMA to prepare annual budget of current accounts at the microsegment level with a MAPE of 12%.

Performed data cleaning/aggregation and designed a loyalty campaign to reduce churn rate by promoting purchase of products positively correlated to customer loyalty. The model, built using logistic regression and random forest, helped reduce attrition/ brand switching by 15% and increased cross-sell penetration by 6%(measured using A/B testing).

Used transactional data to build a customer churn model, using logistic regression and random forest, to predict the churn probability for each customer. The model, with an accuracy of 82%, helped us retain customers worth $15M.

Analyzed the resources data, created ad-hoc reports, designed KPIs for the sales force and developed an analytical framework to increase lead to opportunity conversion rate by 17%.

Designed a marketing campaign to increase customer acquisition and tracked its success using cost-benefit analysis and customer lifetime value. The campaign increased customer acquisition in metropolitan areas by 12%.

ACADEMIC PROJECTS

Marketing Analytics Adstock Marketing-mix Modeling R SQL

Created a marketing-mix model using adstock values of four marketing channels to find the optimal marketing spend for each channel. The model, built using linear optimization and linear regression, helped increase sales by 7.3%.

Risk Analytics Credit card Fraud Detection R

Built a random Forest model to identify fraudulent customers using customer credit card transaction data. The model had a high accuracy and helped the bank inspect fraudulent transactions at real time and prevent a loss of $30k per month.

Customer Analytics The San Francisco Bike Share Visualization R Excel Tableau

Performed data cleansing, data wrangling/munging and designed interactive Tableau dashboards to track KPIs and identify lost opportunities in sales. We suggested differential rental pricing and addition of bikes at the stations to increase the profits by 4%.

Financial Portfolio Analysis Spreadsheet Modeling Advanced Excel

Designed a live stock market portfolio model to analyze gains on investments using Sensitivity analysis and Linear Optimization. The optimal solution, with a ROI of 16.6%, helped us maximize the client’s returns.

Contact this candidate