Lab 1: Exploring and Detecting PII in the PIILO Dataset

Complete the assigned PIILO notebook:

1. Part 1: fill all nine rows of worksheet.your_classification and worksheet.your_justification. Use direct identifier, quasi-identifier, or not PII, and explain each choice in its context.
2. Part 2: run the dataset summaries, baseline detection evaluation, and masking examples. Inspect the results and discuss the detector's limitations.
3. Part 3: fill my_case_original, my_case_masked, and my_case_answer with your own example and a 2–3 sentence explanation of whether the person remains identifiable.

Run all cells, then submit the completed notebook. Use Run all on server or a local Jupyter environment with the packages listed in Setup; the browser runtime may not support all dependencies. The synthetic fallback is sufficient for these checks; identify which data you used when discussing the results.

Automatic completion checks (100 points)

• Nine classifications filled with one of the three categories: 20 points.
• Nine non-empty justifications: 20 points.
• Dataset summaries computed consistently from the loaded data: 15 points.
• Baseline evaluation and masking functions run: 15 points.
• Original and masked examples filled and different: 15 points.
• Anonymity explanation filled: 15 points.

These checks verify completion and execution, not the correctness of written reasoning. The instructor reviews your classifications, justifications, and privacy analysis and may adjust the grade. No minimum detector accuracy is required: analyzing its limitations is part of the lab.