📊 Amazon 2025 Sales Analysis

Dataset: Amazon Sales (250 transactions)
Tools: Python · Pandas · Data Wrangling · Exploratory Data Analysis · Visualization

---

🧠 Project Objective

This project explores sales performance patterns in a 2025 Amazon dataset, focusing on:

• Order status (Completed, Cancelled, Pending)
• Payment methods
• Customer location
• Product categories & pricing

Our goal was to uncover actionable insights that drive sales optimization, customer targeting, and inventory strategy.

---

🔄 Data Preparation

We performed essential data wrangling and cleaning, including:

• Removing duplicates and irrelevant columns (check_total)
• Standardizing messy product names and date formats
• Creating new features like Month and Year for time-based analysis
• Verifying total sales values through a validation column

---

🔍 Hypotheses Explored

Hypothesis ID	Focus Area	Question
H1	Payment Behavior	Do certain payment methods lead to more cancellations?
H2	Location Effect	Do cities influence whether orders get completed or canceled?
H3	Category Performances	Do electronics outperform other categories in total sales?
H4	Product Value	Do high-priced products lead to more completed orders?

---

📈 Key Findings

• 🧾 Gift Cards had the lowest completion rate among all payment methods.
• 🏙️ Miami showed an unexpectedly high number of completed orders.
• 💡 Electronics:
  • Accounted for 63% of completed sales
  • Drove 47.2% of total revenue
  • Had high cancellation rates, revealing volatility
• 🔌 Home Appliances performed strongly in completed sales, just behind Electronics.
• ⚠️ Data integrity issue discovered: total_sales mismatch — solved by introducing a check_total column.

---

🧪 Lessons Learned

• Always validate numeric fields and challenge assumptions.
• Clean, well-structured data leads to faster, deeper insights.
• Visual storytelling is key in translating analytical findings into business strategy.

---

💡 Business Recommendations

• 🎯 Prioritize Electronics in marketing and supply chain planning.
• 🛍️ Offer discounts and incentives for high-value categories.
• 🌍 Target cities with high completion trends (e.g., Miami).
• 📅 Schedule sales around months with highest historical sales.

---

📌 Next Steps

• Analyze customer-level behavior across time
• Run predictive modeling for cancellation likelihood
• Segment by seasonality and product lifecycle

---

👤 Team

Damian Micó Bedoya
Data Analyst | SQL · Python · Tableau | Logistics & Business Intelligence
📫 [email protected] | 📍 Barcelona, Spain
🔗 LinkedIn | GitHub

---

“Clean data, clear insights, smart business.”