BearCart E-commerce Dashboard

Project Overview

This project focused on the critical step of data cleaning and preprocessing. I utilized Python and the Pandas library to build robust pipelines capable of handling large-scale datasets. The primary objective was to ensure data accuracy and consistency, which are foundational for any reliable analysis. The pipelines handled missing values, outlier detection, and data normalization.

Project Gallery

Previous slideNext slide

Key Features & Implementation

• Cleaned and preprocessed large-scale e-commerce datasets including website sessions, orders, products, marketing sources, and refunds using Python (Pandas).
• Recovered missing user identifiers through session-level mapping to preserve customer journey accuracy and validated duplicate transactional records.
• Engineered key business metrics such as gross profit, net revenue, gross margin, refund rate, revenue per session, average order value, and conversion rate.
• Built a multi-page Power BI dashboard featuring executive KPIs, website engagement analysis, device-wise conversion behavior, and funnel performance tracking.
• Performed marketing channel analysis to evaluate traffic quality, conversion efficiency, revenue contribution, and refund impact across sources and campaigns.
• Conducted product-level revenue and profitability analysis to identify high-performing products, margin risks, and refund-driven revenue leakage.
• Used AI-based decomposition trees and Q&A visuals to identify root causes of revenue changes across session type, device, channel, and product dimensions.
• Delivered actionable business recommendations focused on marketing optimization, product prioritization, mobile UX improvement, and refund reduction strategies.