Concepts 🚀 Project 🧮 Labs

EDA visualization — 6 plots that tell you everything

Not 'make a chart'. 6 specific plots, each answering a different question. This is how you SEE the data before modeling.

script.py v3.1

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

# ── Setup: dataset with interesting structure ──
np.random.seed(42)
n = 500
df = pd.DataFrame({
    'age': np.random.normal(40, 12, n).clip(18, 80),
    'income': np.random.lognormal(10.5, 0.6, n),
    'years_experience': np.random.randint(0, 40, n),
    'education': np.random.choice(['High School', 'Bachelor', 'Master', 'PhD'], n),
    'city': np.random.choice(['NYC', 'LA', 'SF', 'Chicago'], n),
})
df['salary'] = (
    30000 + df['years_experience'] * 2500 + df['education'].map({'High School': 0, 'Bachelor': 4000, 'Master': 8000, 'PhD': 12000})
    + np.random.normal(0, 8000, n) + df['city'].map({'NYC': 15000, 'SF': 20000, 'LA': 10000, 'Chicago': 5000})
)

# ════════════════════════════════════════════════════════════════════
#  PLOT 1: HISTOGRAM — what's the SHAPE of each numeric column?
# ════════════════════════════════════════════════════════════════════
fig, axes = plt.subplots(1, 3, figsize=(15, 4))
axes[0].hist(df['age'], bins=30, edgecolor='black', color='#a0c000')
axes[0].set_title('Age (normal distribution)')
axes[0].set_xlabel('Age'); axes[0].set_ylabel('Count')

axes[1].hist(df['income'], bins=30, edgecolor='black', color='#58a6ff')
axes[1].set_title('Income (RIGHT-SKEWED — lognormal)')
axes[1].set_xlabel('Income')

axes[2].hist(np.log(df['income']), bins=30, edgecolor='black', color='#bc8cff')  # log transform
axes[2].set_title('Log(Income) — now normal!')
axes[2].set_xlabel('Log(Income)')
plt.tight_layout()
plt.savefig('plot1_histogram.png', dpi=100, bbox_inches='tight')
plt.show()
print('Plot 1: HISTOGRAM — shows distribution shape')
print('  THINK: income is skewed → log-transform makes it normal → ML models work better')

# ════════════════════════════════════════════════════════════════════
#  PLOT 2: BOX PLOT — outliers + distribution by category
# ════════════════════════════════════════════════════════════════════
fig, axes = plt.subplots(1, 2, figsize=(12, 5))
df.boxplot(column='salary', by='city', ax=axes[0])
axes[0].set_title('Salary by City')
axes[0].set_xlabel('City'); axes[0].set_ylabel('Salary')

df.boxplot(column='salary', by='education', ax=axes[1])
axes[1].set_title('Salary by Education')
axes[1].set_xlabel('Education'); axes[1].set_ylabel('Salary')
plt.tight_layout()
plt.savefig('plot2_boxplot.png', dpi=100, bbox_inches='tight')
plt.show()
print('\nPlot 2: BOX PLOT — outliers (dots) + median + quartiles')
print('  THINK: SF has highest median salary. PhD has highest. Outliers above the whiskers.')

# ════════════════════════════════════════════════════════════════════
#  PLOT 3: SCATTER PLOT — relationship between 2 numeric features
# ════════════════════════════════════════════════════════════════════
fig, ax = plt.subplots(figsize=(8, 5))
colors = {'NYC': '#a0c000', 'LA': '#58a6ff', 'SF': '#f85149', 'Chicago': '#bc8cff'}
for city, color in colors.items():
    subset = df[df['city'] == city]
    ax.scatter(subset['years_experience'], subset['salary'], c=color, label=city, alpha=0.6)
ax.set_xlabel('Years of Experience')
ax.set_ylabel('Salary')
ax.set_title('Salary vs Experience (colored by city)')
ax.legend()
plt.tight_layout()
plt.savefig('plot3_scatter.png', dpi=100, bbox_inches='tight')
plt.show()
print('\nPlot 3: SCATTER — relationship between 2 numeric features')
print('  THINK: positive linear trend (more experience = higher salary).')
print('  THINK: SF cluster is higher than Chicago — city matters.')

# ════════════════════════════════════════════════════════════════════
#  PLOT 4: CORRELATION HEATMAP — which features move together?
# ════════════════════════════════════════════════════════════════════
fig, ax = plt.subplots(figsize=(8, 6))
corr = df.select_dtypes(include=[np.number]).corr()
im = ax.imshow(corr, cmap='RdBu_r', vmin=-1, vmax=1, aspect='auto')
ax.set_xticks(range(len(corr.columns))); ax.set_xticklabels(corr.columns, rotation=45, ha='right')
ax.set_yticks(range(len(corr.columns))); ax.set_yticklabels(corr.columns)
# Add correlation values in each cell
for i in range(len(corr.columns)):
    for j in range(len(corr.columns)):
        ax.text(j, i, f'{corr.iloc[i, j]:.2f}', ha='center', va='center', fontsize=8)
plt.colorbar(im, label='Correlation')
ax.set_title('Correlation Heatmap')
plt.tight_layout()
plt.savefig('plot4_heatmap.png', dpi=100, bbox_inches='tight')
plt.show()
print('\nPlot 4: HEATMAP — correlations at a glance')
print('  THINK: dark red = high positive. Dark blue = high negative.')
print('  THINK: years_experience ↔ salary is red (0.85) — strong predictor.')

# ════════════════════════════════════════════════════════════════════
#  PLOT 5: BAR CHART — categorical distribution
# ════════════════════════════════════════════════════════════════════
fig, axes = plt.subplots(1, 2, figsize=(12, 4))
df['city'].value_counts().plot.bar(ax=axes[0], color='#a0c000')
axes[0].set_title('Count by City'); axes[0].set_ylabel('Count')

df.groupby('education')['salary'].mean().plot.bar(ax=axes[1], color='#58a6ff')
axes[1].set_title('Mean Salary by Education'); axes[1].set_ylabel('Mean Salary')
plt.tight_layout()
plt.savefig('plot5_bar.png', dpi=100, bbox_inches='tight')
plt.show()
print('\nPlot 5: BAR CHART — counts + means per category')
print('  THINK: PhD earns the most. Bachelor > High School. Makes sense.')

# ════════════════════════════════════════════════════════════════════
#  PLOT 6: PAIR PLOT — all pairwise relationships at once
# ════════════════════════════════════════════════════════════════════
from pandas.plotting import scatter_matrix
fig, ax = plt.subplots(figsize=(10, 10))
scatter_matrix(df[['age', 'income', 'years_experience', 'salary']], ax=ax, diagonal='hist', alpha=0.5)
plt.suptitle('Pair Plot — all pairwise relationships')
plt.tight_layout()
plt.savefig('plot6_pairplot.png', dpi=100, bbox_inches='tight')
plt.show()
print('\nPlot 6: PAIR PLOT — scatter for every pair + histogram on diagonal')
print('  THINK: scan for linear patterns (good for linear models).')
print('  THINK: scan for non-linear patterns (need tree models or feature engineering).')
print('\n6 plots = complete EDA. You now KNOW your data.')

Python output click Run to execute

Click "Run" to execute the code on the left.
The output will appear here.

🧠 How it works

6 plots, 6 questions: HISTOGRAM — what's the SHAPE? Normal, skewed, bimodal? Skewed → log-transform. BOX PLOT — outliers + distribution by category. Dots above whiskers = outliers. SCATTER — relationship between 2 numeric features. Linear? Non-linear? Clustered by category? HEATMAP — all correlations at once. Red = positive, blue = negative. Dark = strong. BAR CHART — counts + means per category. Which category is most common? Which has highest target? PAIR PLOT — scatter for EVERY pair + histograms on diagonal. Slow but reveals everything. ORDER MATTERS: histogram first (understand each feature alone), then scatter (pairs), then heatmap (all pairs). Don't skip to modeling until you've made these 6 plots. You'll miss leaks, outliers, and non-linear patterns.

✏️ Try these changes

Change bins=30 to bins=10 in the histogram
→ coarser bins — less detail but clearer shape
Add a log scale to the income histogram: axes[1].set_yscale('log')
→ shows the tail more clearly
Color the scatter by education instead of city
→ different clustering pattern
Add alpha=0.3 to the scatter (more transparent)
→ better for dense data — shows overlap

← Previous lab

EDA mindset — 8 questions that reveal what the data is hiding

Next lab →

Linear Regression maths — the equation behind .fit()