  
• Leading a team to create social data-based products to develop an audience attribution and research platform for major entertainment and media companies. This involved using terabytes of Twitter and Facebook data and creating audience segments from psychographic profiles, TV Shows, Movie and Brand affinities and presenting the insights on a sophisticated GUI for ad analysts and researchers. This was for both publishers, and agencies and advertisers. This was developed on SAS, R and on the Spark ML platform. (Algorithms used were Text analytics, word2vector methods, IBM personality insights, Latent Class Analysis, Logistic Regression and Random Forests)
  
• Developing and operationalizing price elasticity models for a multi-department, multi-product retail store across channels, across time and across major customer segments (Algorithms used were Regression Models, Principal Components Analysis and Latent Class Models; Used Alteryx and R)
  
• Customer lifetime valuation models for the subscribers of a major US cable TV company. This involved combining revenue and costs data and predicting probability of customer churn. The outputs of this model were applied on ~3.2 MM subscribers and used for tailoring and targeting. It was also used for valuing the overall company when it was put up for sale (Algorithms used were Decision Trees, Logistic Regression, Random Forests, Bagging and Boosting; Used SPSS, and SQL)
  
• Global Oil and Gas company: Improve and automate chemometric analysis and model quality evaluation processes for ethanol fermentation for the BioFuels group. Design effective and efficient data acquisition and integration. Create, test and implement the appropriate algorithms. Provide architectural and solution designs. (Algorithms used were Regression, Partial Least Squares and Ensemble Modeling; Neural networks; Used SPSS, SQL and R)
  
• Identify risk factors for hospitalization from Self Monitored Blood Glucose (SMBG) and Electronic Medical Records (EMR) clinical data. This work required extensive data preparation and integration work using SQL and knowledge of ICD codes and CMS rules and regulations. (Algorithm used was Support Vector Machines, SQL and R)

Personal details

Abhijit S.

Summary

Technical skills