Dataset Safety Policy

Applied Machine Learning Algorithms uses synthetic or rights-cleared data only.

Do not use:

• names, emails, phone numbers, addresses, IDs, or account records;
• health, financial, employer, client, legal, security, or authentication data;
• scraped private data;
• production logs; or
• data from a learner's workplace.

Every dataset must have:

• provenance note;
• data dictionary;
• intended prediction-time availability note;
• checksum where practical;
• known limitation note; and
• statement that it is fictional, synthetic, or rights-cleared.

The course may discuss sensitive domains only as cautionary examples. Learners must not build regulated-domain claims from course exercises.