Data Scientist in Digital Business Group - Analytics Team• Implemented an end-to-end Lead Scoring model using GLM for the Motor and Health insurance business, categorizing leads on a scale from 1 to 5.• Utilized GLM (generalized linear model) to achieve a remarkable accuracy of 92% and a kappa value of 0.85.• Increased the conversion rate from 5% to 32%, by identifying and prioritizing hot leads (rated 4 and 5) while eliminating 86% of cold leads (rated 1, 2, and 3).• This optimization resulted in an 86% reduction in efforts, identifying junk leads, and significantly streamlining call center operations.•Utilized ARIMA modeling with parameter tuning (p, d, q) for time series forecasting of leads, NOP, and GWP for FY25, ensuring accurate predictions.•Leveraged time series analysis to anticipate trends and fluctuations, facilitating informed decision-making for budgeting, resource allocation, and financial planning in FY25.• Conducted a cohort analysis on motor and health insurance lines of business, incorporating demographics such as age, gender, population, and city type to enhance targeting strategies.• Applied k-proto clustering to identify age segments and city distribution for policies sold, revealing 6 clusters and patterns for targeted marketing strategies, Additionally, identified the most and least selling policies within each cluster for more effective targeting.

•The objective of the project was to increase sales and reduce losses using machine learning (python) and Power BI.•Performed EDA and developed insights on customer orders, categories & payment methods using Power BI.•Built a classification model to identify fake orders using logistic regression, evaluated the model using a confusion matrix with a model accuracy of 96% & reduced losses by 10%.•Sentiment analysis using NLP (Natural Language Processing) of 1800 reviews with the help of the text blob library to maintain customer retention.

•The objective was to identify the trends and insights into the candy and tobacco market for the Dharampal Satyapal Group.•Designed questionnaires using the Survey-to-go app, carried out quality checks on survey responses to minimize errors by 20% & analyzed consumer shift toward the rival brand.•Monitored survey fieldwork, analyzed 1200 responses for Candy & Tobacco products in Kolkata using Power BI, MS Excel || found that feel, after effect & price, expired products were reasons for the decline in sales respectively.•Presented quantitative data and daily reports to the client in the form of clustered bar graphs, donut charts, etc. which improved my communication skills.