Figure out how to get frequency of names

[CarlKCarlK]

The MemberMatch feature needs, as input, the (approximate) frequency of each member's first name(s) and last names(s). Some alternatives:

• With permission, we could get this from https://www.mynamestats.com/First-Names-Search.html
• With permission, we could get this from any search engine that returns a hit count to web API callers. (Method: Asked the # of hits of John or John phone. Assume "John" has a frequency of 1% (which is about right). Then search Alice or Alice phone and use the relative # of hits to get the approximate frequency for "Alice".)
• There may be some fixed reference (baby names from Social Security?, parsing lots of race results?) we can use. We don't need all names, just the (say) 5,000 most common. After that, we can just assume that anything not in the fixed reference has low frequency. It would be good, however, to have a source that does well for names of people in American born in other countries.

[CarlKCarlK]

Good news -- I've found three sources of free information that I think will give us what we need:

@luillo1, I know little of databases but a lot about processing text CSV files, so how about I write a utility program that will merge these 100's of files into one reasonable CSV. It will have about a few 100K of rows and fewer than a dozen columns. In the short term, MemberMatch can use this one reasonable CSV. In the medium term, you can move it into a database if you want.

• Carl

• https://www.census.gov/topics/population/genealogy/data/2010_surnames.html

• https://www.census.gov/topics/population/genealogy/data/1990_census/1990_census_namefiles.html

• https://catalog.data.gov/dataset/baby-names-from-social-security-card-applications-national-data

[CarlKCarlK]

I saw @satvu yesterday at ESR Track and she said she found these sources, too.

[CarlKCarlK]

I've created a tab-separated file that associates 250K names with their (approximate) probability.

@luillo1 & @MutatedGamer & @satvu The file is 5 meg. Is this OK to check in for now, or is it too big?

I imagine that you'll eventually want it converted to a serialized binary C# dictionary or a database table, but for now it would be useful to have it checked in. It would let me update the MemberMatch functions using this data.