The Python Data Stack

Library Purpose
NumPy Numerical arrays and math
Pandas DataFrames, data cleaning, analysis
Matplotlib / Seaborn Visualization
Jupyter Interactive notebooks
Scikit-learn Machine learning

Installation 

  pip install numpy pandas matplotlib jupyter scikit-learn
jupyter notebook

NumPy Basics 

  import numpy as np

arr = np.array([1, 2, 3, 4, 5])
print(arr.mean())       # 3.0
print(arr * 2)          # [2, 4, 6, 8, 10]

matrix = np.array([[1, 2], [3, 4]])
print(matrix.shape)     # (2, 2)
print(matrix.T)         # transpose

Pandas DataFrames 

  import pandas as pd

df = pd.read_csv('sales.csv')
print(df.head())
print(df.describe())

# Filtering
high_sales = df[df['amount'] > 1000]

# Grouping
by_region = df.groupby('region')['amount'].sum()

# Missing data
df.fillna(0, inplace=True)
df.dropna(subset=['email'], inplace=True)

Visualization 

  import matplotlib.pyplot as plt

df.groupby('month')['revenue'].sum().plot(kind='bar')
plt.title('Monthly Revenue')
plt.xlabel('Month')
plt.ylabel('Revenue ($)')
plt.tight_layout()
plt.savefig('revenue.png')
plt.show()

Scikit-learn Quick Start 

  from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_squared_error

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
model = LinearRegression()
model.fit(X_train, y_train)
predictions = model.predict(X_test)
print(f"MSE: {mean_squared_error(y_test, predictions):.2f}")

Workflow Tips

  1. Explore data with df.info(), df.describe(), visualizations
  2. Clean missing values, outliers, type conversions
  3. Feature engineer new columns from existing data
  4. Model with scikit-learn; evaluate with cross-validation
  5. Document in Jupyter notebooks for reproducibility

Python’s data ecosystem makes it the top choice for analytics, ML, and scientific computing.

Data Cleaning with Pandas 

  import pandas as pd

df = pd.read_csv('sales.csv')
df['date'] = pd.to_datetime(df['date'])
df['amount'] = pd.to_numeric(df['amount'], errors='coerce')
df = df.drop_duplicates(subset=['order_id'])
df = df[df['amount'] > 0]

Merging DataFrames 

  orders = pd.read_csv('orders.csv')
customers = pd.read_csv('customers.csv')

merged = orders.merge(customers, on='customer_id', how='left')

Seaborn for Statistical Plots 

  import seaborn as sns
import matplotlib.pyplot as plt

sns.scatterplot(data=df, x='ad_spend', y='revenue', hue='region')
plt.title('Ad Spend vs Revenue by Region')
plt.savefig('scatter.png', dpi=150)

Train/Test Split Best Practices 

  from sklearn.model_selection import cross_val_score

scores = cross_val_score(model, X, y, cv=5, scoring='accuracy')
print(f"CV accuracy: {scores.mean():.3f} (+/- {scores.std():.3f})")

Never evaluate on training data — use holdout or cross-validation.

Reproducible Notebooks 

  import numpy as np
np.random.seed(42)

Pin library versions in requirements.txt and restart kernel after code changes for consistent results.