Identify and flag high risk customers to mitigate attrition using a proactive, data-based statistical model.
Details: Analyzed and identified features based on subscription, transaction, usages & net promoter score data across
different business units. Developed a python based regression model to categorize existing customers risk levels.
Prepared 4 different models to understand and capture customer behavior and maximize utilization of available data.
Presented data in customizable real time dashboards.
Impact: Forecasted $40M of high risk customers among a portfolio of about $280M value. Real-time dashboards helped
sales person closely monitor customers. Successfully converted pilot project into a full time project with $1.2B portfolio.
Project Scope: Automation of summarized report over competitor development using press release data.
Details: Created python based web scraping tool to import unstructured data from html source into structured CSV
format. Developed python based summarization tool to summarize individual press release content into short consumable
information. Implemented a keyword based model to categorize content into product, merger and financial reports.
Impact: Automated a 4 man team effort into a single click tool generating summarized error-free reports.