Bharat Garg

Developed an RStudio Shiny-based COVID-19 spread simulator to study how various parameters, such as population density, social distancing, and recovery time, would impact the spread of the virus over time.

Developed a random forest model to measure member digital propensity and scored 33 million data entries.
Successfully modeled ranking mechanism basis 172 parameters from demographic, Rx, claims, and orders.
Supervised a team of two, responsible for deliverables, and single-handedly managed all client communications.
Generated key insights and identified the top five positive and negative factors driving digital adoption.
Successfully modeled attrition and achieved a recall of 90% in the top decile with a high digital propensity.

Led a team of three, creating a boosted tree model to measure member propensity to attrite in the next three months. Processed 200 million records from five data sources for 1.5 million members and defined 486 features for modeling.
Achieved a 12% reduction in attrition rate in high propensity group and overall attrition reduced by 3%. Also, I accomplished an increase in response rate by 18% by sending retention campaigns after six months of inactivity.

Developed a tailored inhouse process mining tool using open-source tools and libraries in R and Shiny BI.
Built process discovery, variant explorer, root cause analysis, executive summary all in a single package.

Identified bottlenecks in the existing report design process, established best practices for reporting in Infor BI.
Developed 18 reports using best practices, minimized maintenance time by automating 15+ quality checks.
Achieved a load time reduction of 90% on average and reduced maintenance work full-time equivalent (FTE) from 2 to 0.2.