(Note: I originally published this a few years ago, but realized it’s still relevant today. Enjoy.)
1.Thou shalt decide on a business case before committing any time, money, or animal sacrifice to the project.
2. Thou shalt start small and build, committing not to cataloging the Internet nor to boiling the ocean on thy first project.
3. Thou shalt not spend unnecessarily. Open Source begat Hadoop, and from Open Source did many commercial packages spring forth.
4. Thou shalt not fear the Unstructured Data, for Hadoop and NoSQL are with you.
5. Be fruitful and multiply thy commodity hardware, for dedicated, proprietary hardware is neither holy nor necessary.
6. Thou shalt speak in tongues, like those named Python, SQL and Java.
7. IT shalt not wage war on thy brethren the Businessmen, for without them thou hast no problem to solve, and without you they have no solution.
8. Thou shalt not use the unclean data. Thou must ritually cleanse it first, ensuring the salvation of thine project thereby.
9. Thou shalt not overfit the model, lest thou burn in statistical hell forevermore.
10. Thou shalt trust thy results, but verify with rituals such as the Confusion Matrix.