Roteiro Python para ciência de dados 2026: pandas, scikit-learn e PyTorch

import numpy as np

# NumPy is the foundation
arr = np.array([1, 2, 3, 4, 5])
print(arr * 2)           # [2, 4, 6, 8, 10]
print(arr.mean())        # 3.0
print(arr.std())         # 1.41...

# Matrix operations
matrix = np.random.randn(5, 5)
print(matrix.shape)      # (5, 5)
print(matrix.sum(axis=0)) # column sums

import pandas as pd

# Load and explore
df = pd.read_csv("sales.csv")
print(df.shape)           # (rows, columns)
print(df.dtypes)          # column types
print(df.describe())      # statistics
print(df.isnull().sum())  # missing values

# Clean
df = df.dropna(subset=['price'])  # drop rows with missing price
df['date'] = pd.to_datetime(df['date'])
df['category'] = df['category'].str.strip().str.lower()

# Analyze
monthly = df.groupby(df['date'].dt.month)['revenue'].sum()
top_products = df.groupby('product')['quantity'].sum().sort_values(ascending=False).head(10)

# Merge datasets
customers = pd.read_csv("customers.csv")
merged = df.merge(customers, on='customer_id', how='left')

import matplotlib.pyplot as plt
import seaborn as sns
import plotly.express as px

# Publication-quality static charts (Matplotlib/Seaborn)
fig, axes = plt.subplots(2, 2, figsize=(12, 8))
axes[0,0].hist(df['revenue'], bins=30, color='steelblue')
axes[0,0].set_title('Revenue Distribution')
sns.boxplot(data=df, x='category', y='revenue', ax=axes[0,1])
plt.tight_layout()
plt.savefig('analysis.png', dpi=300)

# Interactive charts (Plotly)
fig = px.scatter(df, x='marketing_spend', y='revenue',
                 color='category', size='quantity',
                 hover_name='product', trendline='ols')
fig.show()  # opens in browser

from sklearn.model_selection import train_test_split, cross_val_score
from sklearn.preprocessing import StandardScaler
from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor
from sklearn.metrics import mean_absolute_error, r2_score
import joblib

# Prepare data
X = df[['marketing_spend', 'season', 'product_category', 'competitor_price']]
y = df['revenue']

# Encode categoricals
X = pd.get_dummies(X, columns=['season', 'product_category'])

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train)
X_test_scaled = scaler.transform(X_test)

# Train and evaluate multiple models
models = {
    'Random Forest': RandomForestRegressor(n_estimators=200, random_state=42),
    'Gradient Boosting': GradientBoostingRegressor(n_estimators=200, random_state=42),
}

for name, model in models.items():
    model.fit(X_train_scaled, y_train)
    y_pred = model.predict(X_test_scaled)
    mae = mean_absolute_error(y_test, y_pred)
    r2 = r2_score(y_test, y_pred)
    cv = cross_val_score(model, X_train_scaled, y_train, cv=5, scoring='r2')
    print(f"{name}: MAE={mae:.0f}, R2={r2:.3f}, CV={cv.mean():.3f}±{cv.std():.3f}")

# Save best model
joblib.dump(models['Random Forest'], 'revenue_model.joblib')

import torch
import torch.nn as nn
from torch.utils.data import DataLoader, TensorDataset

# Simple neural network for tabular data
class SalesNet(nn.Module):
    def __init__(self, input_dim: int):
        super().__init__()
        self.net = nn.Sequential(
            nn.Linear(input_dim, 256),
            nn.ReLU(),
            nn.Dropout(0.3),
            nn.Linear(256, 128),
            nn.ReLU(),
            nn.Dropout(0.2),
            nn.Linear(128, 1),
        )

    def forward(self, x):
        return self.net(x).squeeze()

# Training loop
model = SalesNet(input_dim=X_train.shape[1])
optimizer = torch.optim.Adam(model.parameters(), lr=1e-3)
criterion = nn.MSELoss()

for epoch in range(100):
    model.train()
    optimizer.zero_grad()
    preds = model(X_tensor)
    loss = criterion(preds, y_tensor)
    loss.backward()
    optimizer.step()