lab solved

data/raw/lab_eda_univariate_insitghs.md

@@ -0,0 +1,179 @@
+# 📊 Amazon UK Product Analysis – Univariate EDA
+
+**By:** Paulina Mamiaga  
+
+---
+
+## 📌 Objective
+
+The objective of this analysis is to explore and understand the structure and distribution of product listings on Amazon UK through univariate exploratory data analysis (EDA).
+
+This study focuses on key product attributes such as **category, price, and customer ratings** in order to:
+
+- Identify the most represented product categories and assess market concentration  
+- Analyze pricing patterns, including central tendencies and variability  
+- Understand customer rating behavior and detect potential biases or trends  
+
+From a business perspective, this analysis aims to generate insights that support **product positioning, pricing strategies, and inventory decisions** in a competitive e-commerce environment.
+
+---
+
+## 📂 Data Source
+
+Dataset used:  
+👉 https://www.kaggle.com/datasets/asaniczka/uk-optimal-product-price-prediction/
+
+---
+
+# 🧠 Part 1: Understanding Product Categories
+
+## 🔍 Key Findings
+
+- The dataset shows **extreme concentration in a single category: _Sports & Outdoors_**
+- This category dominates the dataset, representing the vast majority of listings
+- Other categories (Beauty, Bath & Body, etc.) contribute only marginal shares (~2% each)
+
+## 💡 Business Insights
+
+👉 **Market Concentration Risk**
+- The dataset is highly skewed → global insights may be misleading  
+- Results are heavily driven by one dominant category  
+
+👉 **Competitive Dynamics**
+- High concentration → high competition  
+- Smaller categories → potential niche opportunities  
+
+👉 **Recommendation**
+- Always segment analysis by category  
+- Identify underrepresented categories for strategic positioning  
+
+---
+
+# 💰 Part 2: Product Pricing Analysis
+
+## 📊 Central Tendency
+
+- Mean price: **£89.24**
+- Median price: **£19.09**
+- Mode price: **£9.99**
+
+## 📊 Dispersion
+
+- Standard deviation: **345.61**
+- Range: **£0 – £100,000**
+- IQR: **£36**
+
+## 🔍 Key Findings
+
+- Strong gap between mean and median → **right-skewed distribution**
+- Majority of products are low-priced
+- Presence of extreme high-value outliers
+
+## 💡 Business Insights
+
+👉 **Skewed Pricing Structure**
+- Marketplace operates in a **low-price, high-volume model**
+
+👉 **Mean is Misleading**
+- Average price is inflated by outliers  
+- Median better represents typical product pricing  
+
+👉 **Outliers**
+- Extreme values distort analysis  
+- Likely luxury or incorrectly priced products  
+
+👉 **Recommendation**
+- Use median & IQR for pricing strategy  
+- Segment pricing by category  
+- Consider removing or capping outliers  
+
+---
+
+# ⭐ Part 3: Product Ratings Analysis
+
+## 📊 Centrality
+
+- Mean rating: **2.15**
+- Median rating: **0.0**
+- Mode rating: **0.0**
+
+## 📊 Dispersion
+
+- Standard deviation: **2.19**
+- IQR: **4.4**
+
+## 🔍 Key Findings
+
+- 50% of products have rating = 0  
+- Sharp jump from 0 → 4.4 (75th percentile)  
+- Ratings are not normally distributed  
+
+## 💡 Business Insights
+
+👉 **Critical Interpretation**
+- Rating = 0 likely means **no reviews**, not poor quality  
+
+👉 **Hidden Structure**
+Two groups exist:
+1. Products without ratings  
+2. Products with high ratings (4–5 stars)  
+
+👉 **Misleading Average**
+- Mean rating underestimates real customer satisfaction  
+
+👉 **Business Opportunity**
+- Many products lack reviews → opportunity to influence perception early  
+
+👉 **Recommendation**
+- Separate analysis:
+  - Rated products (stars > 0)
+  - Unrated products (stars = 0)  
+
+---
+
+# 🚀 Overall Business Insights
+
+The Amazon UK marketplace shows strong structural imbalances:
+
+### 1. Category Concentration
+- One category dominates the dataset  
+- Risk of biased insights  
+
+### 2. Price Skewness
+- Most products are low-priced  
+- Few outliers distort averages  
+
+### 3. Rating Bias
+- Many products have no ratings  
+- Customer satisfaction is higher than it appears  
+
+---
+
+## 🎯 Final Takeaways
+
+- Always segment before analyzing  
+- Median > Mean for pricing decisions  
+- Ratings must be interpreted carefully  
+- Unrated products = growth opportunity  
+
+---
+
+## 🛠️ Tech Stack
+
+- Python (Pandas, NumPy)
+- Matplotlib / Seaborn
+- Jupyter Notebook  
+
+---
+
+## 📎 Notes
+
+- No missing values detected  
+- Duplicate products exist at ASIN level  
+- Large dataset (~2.4M rows) ensures robustness  
+
+------
+
+## 🌐 Data Source
+
+[https://www.kaggle.com/datasets/asaniczka/uk-optimal-product-price-prediction/]