Business Intelligence Network

Blog: Pete Loshin

« Using Eclipse for Development, and Web 2.0 Rave | Main | Instructional Eclipse/Java Videos »

Fake Names for Realistic Data

I love this: Fake Name Generator.

My first thought: what a great way to generate some real-sounding but fake names for writers who are having trouble coming up with good names for their characters. Not every writer has the imagination of, say, Charles Dickens. I mean, who else could come up with names like Martin Chuzzlewit, Thomas Gradgrind, Barnaby Rudge and the Marquis de St Evremonde? Check out more cool Dickensian handles at the Dickens Characters page.

Of course, there's more to the Fake Name Generator than that. You don't just get fake names, you get full personal information including fake phone numbers, email addresses, birth dates, Social Security numbers, home address and even mother's maiden names.

If you "order" in bulk (still free, unless you want faster delivery), you can specify things like data format (MS SQL or MySQL, tab-delimited, HTML tables, etc), as well as get even more fake data like credit card numbers and expiration dates.

So, what's better than getting fake random names? Getting fully authentic looking datasets of fake random names, including all the critical data that identity thieves want. Therefore, my second thought: How long before some enterprising soul start putting together complete fake datasets and tries selling them to bottom-feeding mass marketer/spammer/telemarketing firms?

Or, you could just generate some nice-looking datasets for demos, testing, what have you.

  Posted by Pete Loshin on July 27, 2007 8:00 AM |

Comments

Of course, there is a good algorithm for fake name generation in Dav Pilkey's epic novel, "Captain Underpants and the Perilous Plot of Professor Poopypants."

Posted by: David Loshin | July 31, 2007 6:34 AM

Actually, the algorithm you cite (a.k.a. "Name Change-O-Chart 2000") isn't as useful for generating fake names as it is for collapsing the entire human namespace into a set of only 26^^3 (17,576) distinct names.

Pilkey's system comprises three lists of names/parts of names that correspond to the letters of the alphabet, and it actually collapses the namespace rather than generate fake names.

The first list is for first names: a given name is replaced in this system by one of 26 names. Any name starting with the letter "A" is replaced by "Stinky", names starting with "G" are replaced by "Fluffy", and so on. Thus, Alexander, Anne, Abigail, Arthur and Archimedes all become "Stinky"; George, Greta, Gulliver, Gigi, and Gustave all become "Fluffy". All conceivable given names are collapsed into a set of 26 names.

The second and third lists are associated with the first and last letters of the last name: David and I share the surname "Liverchunks" under this system with anyone whose last name starts with "L" and ends in "N": thus the names David Loshin, "Daniel Livingston", "Daria Lin", and "Drusilla Lebron" would all become "Gidget Liverchunks".

All surnames, therefore, are collapsed down to a set of 676 possible names.

Just saying.

Posted by: Pete Loshin | July 31, 2007 8:18 AM

Post a comment