Monday, October 17, 2016

A Linguist Who Cracks the Code in Names to Predict Ethnicity - Note for a Discussion, "E Pluribus Unum? What Keeps the United States United."


As told to PERRY GARFINKEL OCT. 15, 2016, New York Times

Image from entry, with caption: Lisa Spira, 30, is director of research and product development at Ethnic
Technologies in South Hackensack, N.J.

Q. What is your educational background?

A. My degree from Syracuse University was in linguistics. I studied a branch of
linguistics called onomastics, which involves the history and origin of proper names.
In my field I am known as an onomastician.

Have you always been fascinated with names?

Yes. After my sister was born, I took my parents’ baby­naming book and never
gave it back. I even added new names. I had Playmobil toys with about 150 plastic
people. Not only did I name every one, but I also gave them name tags.

What do you do at Ethnic Technologies?

I lead a team that develops our software that predicts individuals’ ethnic origins
based on their full names, addresses and ZIP codes. We build predictive algorithms
based on patterns in names from various ethnic groups. We also track demographic
data that pinpoints ethnic breakdowns by geography. We identify 158 distinct
ethnicities, with further segmentation for Hispanics and African-­Americans.

Can you give an example of how your company’s software works?

Let’s hypothetically take the name of an American: Yeimary Moran. We see the
common name Mary inside her first name, but unlike the name Rosemary, for
example, we know that the letter string “eimary” is Hispanic. Her surname could be
Irish or Hispanic. So then we look at where our Yeimary Moran lives, which is
Miami. From our software, we discover that her neighborhood is more Hispanic
than Irish. Customer testing and feedback show that our software is over 90 percent
accurate in most ethnicities, so we can safely deduce that this Yeimary Moran is

What types of companies come to you for your services?

Any company that wants to target its goods or services to a particular ethnic
group. A perfect example is cosmetics. African­-Americans, Asians, Hispanics and
Caucasians may prefer different cosmetics.

What if companies want to target ethnic groups for the wrong

We vet every potential client to make sure it is reputable. We work hard to make
sure clients don’t have suspicious motivations for using this information. Our
contracts specify allowable uses.

No comments: