Free build-along · Fictional business example

Can a simple Python script become a service someone pays for?

Use this CSV cleanup project to learn the workflow, frame a narrow offer, and test whether a real customer wants it. The example code is free. No email is required.

1. Run the free workflow

Download the self-contained Python script and fictional input file. It uses only the standard library and never uploads anything to a server.

Download Python script Download sample CSV

python csv-cleanup-demo.py sample-orders.csv output

Expected terminal result: Accepted: 2; needs review: 3. The output directory will contain accepted.csv and review.csv. Running it again produces the same files. Read the script to see how it handles duplicate IDs, missing fields, and invalid totals.

Understand the checks

The script trims fields, lowercases email addresses, accepts the first valid occurrence of an order ID, and sends later duplicates to review. Amounts must be nonnegative whole cents. The email check catches only obvious formatting problems; it does not verify an address exists. Review files may contain untrusted spreadsheet formulas: inspect them as plain text, not by opening them directly in Excel.

Troubleshooting and practice

Use Python 3.10 or newer. If the input is not found, run the command in the download folder or provide its full path. A column error means the file does not match this example schema. Each run replaces the two output files; keep an original backup and choose a separate output folder.

Exercise: add a unique order with total_cents set to 0. Expect three accepted records. Then change that amount to 12.50: it should move to review. Compare both output files, not just the terminal counts. See the official csv documentation.

2. Turn the technical result into a narrowly defined offer

3. Validate the offer before building more

Speak with a few relevant people, ask how they currently detect bad rows, and offer to demonstrate a synthetic or properly authorized sample. Confirm their expected outputs, pricing constraints, data permissions, and what would count as a successful paid pilot. Do not call the project a business success until someone actually buys it.

4. Put your work in front of people organically

Publish a helpful demonstration for a precise search query, share it where its intended readers actively ask related questions, and invite useful feedback—not unsolicited marketing. Link your tutorial to your portfolio or a relevant offer, then measure qualified conversations and actual purchases separately.

You can also study the CSV cleaning guide, reading and writing CSV in Python, or the free learning hub.