Names Datasets

Name data for people born in the US, Spain, Argentina, Chile and Italy — collected from five national sources, cleaned, and normalized into one consistent format.

On Kaggle: https://www.kaggle.com/datasets/rosinascampino/name-datasets-us-spain-argentina-chile-italy

The problem

Every country publishes this data its own way. Some offer clean CSV downloads, some publish Excel workbooks, and some only render HTML tables on a government site. Column names, encodings, date ranges and structures don't agree with each other, and none of them were designed to be combined.

I started this as a way to learn data cleaning in Python, and most of the work turned out to be exactly that mismatch: deciding on one target schema and writing a separate path to it from each source.

How it's organized

raw_data/      untouched files as downloaded or scraped
scripts/       collection and cleaning
cleaned_data/  normalized output
sources.txt    where each dataset came from

Keeping raw inputs separate from transformed outputs means the whole thing can be re-run and checked. If a cleaned file looks wrong, the original is still there to compare against, and sources.txt says where it came from.

What I learned

• Reading and writing CSV with Python's csv module
• Web scraping with Beautiful Soup
• Working with Excel files via openpyxl
• Creating and instantiating data classes
• Type hints