In Module 2, we wrote a simple, centralized data_prep.py to get started. In real projects, this often evolves into modular per-table scripts to better manage complexity and focus. We'll see an example of that kind of evolution.

In Module 2, we had one file: scripts/data_prep.py.

In Module 3, we now use one file per data table:

• scripts/data_prep/prepare_customers.py
• scripts/data_prep/prepare_products.py
• scripts/data_prep/prepare_sales.py

As data projects grow, it becomes easier to:

• Focus on one dataset at a time
• Avoid breaking other code when cleaning changes
• Test and debug more easily
• Let different team members work on different files

We move the old data_prep.py in an archive/ folder so you can compare and reuse as needed.

We don't need to create our .venv as we should already have it. If not, go back to Module 1 and 2 make sure those steps are completed. Now, we just follow our regular workflow. If we find we need additional external packages, we can always re-run the install from requirements.txt command as needed. In general, we:

1. Pull any recent changes from GitHub.
2. Activate the .venv.
3. Run scripts/data_prep.py.

Open your smart sales repository in VS Code. Open a terminal in the root project folder. Activate your .venv and run each file.

source .venv/bin/activate
python3 scripts/data_prep/prepare_customers.py
python3 scripts/data_prep/prepare_products.py
python3 scripts/data_prep/prepare_sales.py

Open your smart sales repository in VS Code. Open a PowerShell terminal in the root project folder. Activate your .venv and run each file.

.venv\Scripts\activate
py scripts/data_prep/prepare_customers.py
py scripts/data_prep/prepare_products.py
py scripts/data_prep/prepare_sales.py

Once you’ve verified the scripts ran successfully, git add, commit, and push changes to your GitHub repository.

git add .
git commit -m "ran initial data_prep.py"
git push -u origin main

For best results, git add-commit-push frequently after making any useful progress.

For this step, use pandas (and optionally, a shared DataScrubber class if you like) to clean and prepare each of the raw data files.

Cleaning is a critical task.

Continue until you think you have good data in all the prepared files.

• pro-analytics-01
• And the repos from earlier modules.