Why Every Data Analyst Should Learn Docker in 2026

# Dockerfile for a reproducible analysis
FROM python:3.12-slim
WORKDIR /work
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
COPY report.py .
CMD ["python", "report.py"]

# requirements.txt (pin exact versions you tested with), e.g.:
#   pandas==2.2.2
#   openpyxl==3.1.5

# Build and run, mounting a data folder:
#   docker build -t sales-report .
#   docker run --rm -v "$PWD/data:/work/data" sales-report

from pathlib import Path
import pandas as pd

data_dir = Path("data")
data_dir.mkdir(exist_ok=True)
src = data_dir / "sales.csv"
if not src.exists():   # sample data so the example runs
    pd.DataFrame({"region": ["North", "South", "North"], "amount": [500, 300, 450]}).to_csv(src, index=False)

df = pd.read_csv(src)
summary = df.groupby("region", as_index=False)["amount"].sum()
summary.to_csv(data_dir / "summary.csv", index=False)
print("pandas", pd.__version__)
print(summary)

#!/usr/bin/env bash
set -euo pipefail
mkdir -p init
cat > init/01_sales.sql <<'SQL'
CREATE TABLE sales (id SERIAL PRIMARY KEY, region TEXT, sale_date DATE, amount NUMERIC);
INSERT INTO sales (region, sale_date, amount) VALUES
 ('North', '2026-01-05', 500), ('South', '2026-01-09', 300),
 ('North', '2026-02-02', 450), ('South', '2026-02-11', 380);
SQL
docker run -d --name sqlpractice -e POSTGRES_PASSWORD=practice \
  -p 5432:5432 -v "$PWD/init:/docker-entrypoint-initdb.d" postgres:16
sleep 5
docker exec -it sqlpractice psql -U postgres -c "SELECT region, SUM(amount) FROM sales GROUP BY region;"

-- Monthly sales with month-over-month change
WITH monthly AS (
  SELECT DATE_TRUNC('month', sale_date) AS month, SUM(amount) AS total
  FROM sales
  GROUP BY 1
)
SELECT month,
       total,
       total - LAG(total) OVER (ORDER BY month) AS change
FROM monthly
ORDER BY month;

-- Clean up when finished: docker rm -f sqlpractice

services:
  db:
    image: postgres:16
    environment:
      POSTGRES_PASSWORD: practice
    volumes:
      - ./init:/docker-entrypoint-initdb.d
    ports: ["5432:5432"]

  notebook:
    image: jupyter/scipy-notebook:2024-05-27
    ports: ["8888:8888"]
    volumes:
      - ./notebooks:/home/jovyan/work
    environment:
      DATABASE_URL: postgresql://postgres:practice@db:5432/postgres
    depends_on: [db]

# Run: docker compose up
# Open the Jupyter link shown in the logs.

# Inside the notebook container
# (install the driver if the image lacks it: %pip install psycopg2-binary sqlalchemy)
import os
import pandas as pd
from sqlalchemy import create_engine

engine = create_engine(os.environ["DATABASE_URL"])
df = pd.read_sql("SELECT region, SUM(amount) AS total FROM sales GROUP BY region", engine)
df

Project: Reproducible sales analysis

- PostgreSQL with sample data loaded by init scripts
- Jupyter notebook: SQL queries + pandas analysis + charts
- Report script exporting summary.csv and an Excel file
- Dockerfile with pinned library versions
- compose.yaml starting everything with one command
- README: 'docker compose up', what you will see, findings

Week 1  Docker basics: images, containers, volumes, ports
Week 2  PostgreSQL in Docker; load data; practise SQL
Week 3  Dockerfile for your report; Compose with DB + notebook; README

## Run this project

Requirements: Docker Desktop (or Docker Engine) with Compose.

    docker compose up

Then open the Jupyter link printed in the terminal and run
`notebooks/sales_analysis.ipynb`.

To stop and remove everything:

    docker compose down -v