Summary Statistics

import pandas as pd
import numpy as np

# Load the ecommerce data
df = pd.read_csv('dataplexa_ecommerce.csv')

# First look at our revenue data
print("Sample revenue data:")
print(df['revenue'].head(10))

# Calculate all three measures of central tendency
mean_revenue = df['revenue'].mean()
median_revenue = df['revenue'].median()
mode_revenue = df['revenue'].mode()[0]  # mode() returns a Series, take first value

print(f"Mean revenue: ₹{mean_revenue:,.2f}")
print(f"Median revenue: ₹{median_revenue:,.2f}")
print(f"Mode revenue: ₹{mode_revenue:,.2f}")

# Show the difference between mean and median
difference = mean_revenue - median_revenue
print(f"\nDifference (Mean - Median): ₹{difference:,.2f}")

# Calculate measures of variability
revenue_std = df['revenue'].std()
revenue_var = df['revenue'].var()
revenue_range = df['revenue'].max() - df['revenue'].min()

print("=== VARIABILITY MEASURES ===")
print(f"Standard Deviation: ₹{revenue_std:,.2f}")
print(f"Variance: ₹{revenue_var:,.2f}")
print(f"Range: ₹{revenue_range:,.2f}")
print(f"Min: ₹{df['revenue'].min():,.2f}")
print(f"Max: ₹{df['revenue'].max():,.2f}")

# Apply the 68-95-99.7 rule (assuming normal distribution)
mean_rev = df['revenue'].mean()
std_rev = df['revenue'].std()

# Calculate the ranges
one_std = (mean_rev - std_rev, mean_rev + std_rev)
two_std = (mean_rev - 2*std_rev, mean_rev + 2*std_rev)
three_std = (mean_rev - 3*std_rev, mean_rev + 3*std_rev)

print("68% of orders fall between:", f"₹{one_std[0]:,.0f} - ₹{one_std[1]:,.0f}")
print("95% of orders fall between:", f"₹{two_std[0]:,.0f} - ₹{two_std[1]:,.0f}")
print("99.7% of orders fall between:", f"₹{three_std[0]:,.0f} - ₹{three_std[1]:,.0f}")

# Check actual percentages in our data
within_1std = df[(df['revenue'] >= one_std[0]) & (df['revenue'] <= one_std[1])]
actual_pct = len(within_1std) / len(df) * 100
print(f"\nActual % within 1 std dev: {actual_pct:.1f}%")

# Calculate key percentiles and quartiles
percentiles = [5, 10, 25, 50, 75, 90, 95]
results = df['revenue'].quantile([p/100 for p in percentiles])

print("=== PERCENTILE ANALYSIS ===")
for i, p in enumerate(percentiles):
    print(f"{p:2}th percentile: ₹{results.iloc[i]:,.2f}")

# Calculate Interquartile Range (IQR)
Q1 = df['revenue'].quantile(0.25)
Q3 = df['revenue'].quantile(0.75)
IQR = Q3 - Q1

print(f"\nInterquartile Range (Q3-Q1): ₹{IQR:,.2f}")
print("This captures the middle 50% of your revenue distribution")

# Calculate skewness and kurtosis for the overall dataset
from scipy import stats

revenue_skew = df['revenue'].skew()
revenue_kurt = df['revenue'].kurtosis()

print("=== DISTRIBUTION SHAPE ===")
print(f"Skewness: {revenue_skew:.3f}")
print(f"Kurtosis: {revenue_kurt:.3f}")

# Interpret the results
if revenue_skew > 0.5:
    skew_interpretation = "Right-skewed (long tail of high values)"
elif revenue_skew < -0.5:
    skew_interpretation = "Left-skewed (long tail of low values)"
else:
    skew_interpretation = "Approximately symmetric"

print(f"Interpretation: {skew_interpretation}")

# Compare distribution shapes across product categories
category_stats = df.groupby('product_category')['revenue'].agg([
    'mean', 'median', 'std', 'skew'
]).round(2)

# Add a "mean vs median" ratio to quickly spot skewness
category_stats['mean_median_ratio'] = (category_stats['mean'] / category_stats['median']).round(2)

print("Distribution comparison by category:")
print(category_stats)

# Step 1: Quick overview with describe()
print("=== STEP 1: QUICK OVERVIEW ===")
print(df['rating'].describe())
print("\n" + "="*40)

# Step 2: Check for data quality issues
print("=== STEP 2: DATA QUALITY CHECK ===")
print(f"Missing values: {df['rating'].isnull().sum()}")
print(f"Duplicate values: {df['rating'].duplicated().sum()}")
print(f"Values outside 1-5 range: {((df['rating'] < 1) | (df['rating'] > 5)).sum()}")

# Step 3: Distribution shape analysis
print(f"\n=== STEP 3: DISTRIBUTION SHAPE ===")
print(f"Skewness: {df['rating'].skew():.3f}")
print(f"Mean vs Median gap: {df['rating'].mean() - df['rating'].median():.3f}")
print(f"Coefficient of Variation: {df['rating'].std() / df['rating'].mean():.3f}")

if df['rating'].mean() > df['rating'].median():
    print("→ Right-skewed: Few very high ratings")
else:
    print("→ Left-skewed: Few very low ratings")

# Step 4: Outlier detection using IQR method
Q1 = df['rating'].quantile(0.25)
Q3 = df['rating'].quantile(0.75)
IQR = Q3 - Q1

lower_bound = Q1 - 1.5 * IQR
upper_bound = Q3 + 1.5 * IQR

outliers = df[(df['rating'] < lower_bound) | (df['rating'] > upper_bound)]

print(f"=== STEP 4: OUTLIER DETECTION ===")
print(f"IQR bounds: {lower_bound:.2f} to {upper_bound:.2f}")
print(f"Outliers found: {len(outliers)} ({len(outliers)/len(df)*100:.1f}%)")

# Step 5: Business context interpretation
print(f"\n=== STEP 5: BUSINESS INSIGHTS ===")
print(f"Average rating of {df['rating'].mean():.2f} suggests moderate satisfaction")
print(f"Standard deviation of {df['rating'].std():.2f} shows mixed opinions")

# Most common rating
mode_rating = df['rating'].mode()[0]
mode_count = (df['rating'] == mode_rating).sum()
print(f"Most common rating: {mode_rating} ({mode_count/len(df)*100:.1f}% of orders)")