Arunima Sarkar

Working as a part of Consumer Goods & Services team where I have given my strong contribution to the following projects.

• Employee Performance & Employee Attrition: Used Logistic Regression, Random Forest and XGBoost algorithms to build models which helped in identifying the attrition given a set of features of the employees. Also used topic modeling and sentiment analysis to understand the main reasons behind attrition. Additionally, used NLP on unstructured data to generate word clouds (bigrams and trigrams) to see the performance of the leadership team visually. Tool: Python, PowerBI
• Promotion Analytics (RGM): Built sales decomposition models using Unobserved Component Model (UCM) technique for a beverage company to decompose sources of volume into multiple components and evaluate ROI and uplift for discount and promotion mechanics for sell-in and sell-out data. Also, worked on creating a personalized promotion calendar for entire portfolio by optimization and simulation of mechanics using Genetic Algorithm and MILP for optimization. This derived business a revenue of ~1M USD and reduced manual effort by 30%. Tool: Python, Power BI
• Marketing Mix Modeling (MMM): Built Marketing Mix Models for a brewery giant by automating the entire process starting from data processing, modeling to post modeling result summarization to obtain ROI_s of various marketing channels (traditional & digital). Mixed and Random Effects regression technique has been used to obtain the contribution of base, trade and media variables across countries, brands throughout the year. Tool: Python, Power BI
• Category Forecasting: Built a live app leveraging Linear Regression models on Azure environment to understand the base and covid impact on sales volume for different categories for a Fortune 500 cleaning company. This app helps users to understand the forecasted sales volume of different categories by simulating across different business scenarios. Docker based deployment has been used here. Tool: Python, Streamlit
• Data Products: Built an internal app to understand and overcome challenges faced by any CPG company when dealing with multiple data sets. This app takes all the required data from users, does the mapping of the columns using LLM and after approval sends it to the database. Once we have all the required data, it helps to create an ERD. Further, it provides all kinds of insights for the data sets using Gen AI and does data quality checks typically done before start of major work. Tools: Python, Streamlit
• Formerly Bridgei2i Analytics Solutions

Worked as part of pricing team where I have analyzed pricing and discounting for APJ countries.

• Technology Services Segmentation: Used Segmentation approach to group customers based on their care pack buying behavior and identify opportunity area to upsell and increase revenue. K Means Segmentation has been used to break the customers into different clusters. Also, Recency Frequency Monetary (RFM) approach has been used to shortlist the customers and come up with a list of customers who could be targeted for upselling higher care pack. Tool: R
• Market Basket Analysis: Used Market Basket Analysis to recommend a set of Networking Products. Transaction Data for a time span of two years was chosen and corresponding metrics like Support, Confidence and Lift were calculated to give the recommendation. Apriori Algorithm was mainly used. Additionally cross sell was implemented based on business rules. Tool: Python