David's Computer Stuff Journal comments

"Neat Hecl app from newcomer" by Lauri Ojansivu

Fri, 16 May 2008 16:25:11 -0700

You're welcome! :)

I found about Hecl from wiki.tcl.tk , and with your and other Hecl community help got it running.

Thanks to all for creating this nice language!

"Restaurants, immigrants, and the popularity of various cuisines" by M.Pomme

Fri, 16 May 2008 00:11:24 -0700

Hello Dave,

OK send it along please :)

"Restaurants, immigrants, and the popularity of various cuisines" by Dave Welton

Wed, 14 May 2008 12:56:18 -0700

M, I'd be happy to give you the original spreadsheet if you want to fiddle with and improve on it. This was just a fun exercise for me, though, so I don't think I'm really likely to sink that much more effort into it.

BTW, the numbers in my chart are different from the raw yahoo hits. There are tons of hits for things like Ristorante Cinese, but that's also balanced out by a great deal of Chinese immigrants, for example.

"Restaurants, immigrants, and the popularity of various cuisines" by Seth Roberts

Tue, 13 May 2008 17:26:23 -0700

Thanks for the publicity. You might want to take logarithms before plotting your data. It might make it easier to see a pattern.

"Restaurants, immigrants, and the popularity of various cuisines" by M.Pomme

Tue, 13 May 2008 15:25:35 -0700

OK, maybe I was a little harsh...

It would be interesting to compare this graph with the graph you obtain with just the number of google hits (not "normalized" in any way by the immigration office data). Does it look very different ?

Alternatively could you try using the data to find a estimation of a relation between the popularity and the population:

e.g., if we assume that there is such a relation and that it looks like this:

of google hits (G) = K * (population (P))^n

We can find n by choosing it so that the plot of G/P^n against P is the closest to its linear regression, and K is the coefficient for the linear regression.

Then, just make a bar graph of G/(K*P^n). The food from the country which value is > 1 tastes inherently "better" than the average food, and the ones with values < 1 tastes inherently worse than average :D

He he thanks for this post, I had forgotten maths could be fun...

I hope the theory above is not far too ridiculous ;)

"Restaurants, immigrants, and the popularity of various cuisines" by Dave Welton

Tue, 13 May 2008 11:57:20 -0700

I don't think they're "meaningless" numbers. I think they're rough, imperfect numbers with some potential inaccuracies, but I think they do confirm what 1) I've seen around town in a place like Padova, 2) what people I know like. There are doubtless some outliers (Mongolo, probably also due to its other meaning in Italy), but still, I think there's some truth there.

"Restaurants, immigrants, and the popularity of various cuisines" by M.Pomme

Tue, 13 May 2008 09:54:21 -0700

Just a comment on this:

"However, it seems reasonable that people would mostly talk about restaurants in proportion to their popularity"

While it sounds reasonable to assume people talk more about restaurants which are more popular, it's quite a leap to assume that the correlation is linear.

My uninformed guess would be an exponential correlation (or at least polynomial: x^2, x^3, x^n...), because talk generates more talk (viral marketing and "buzz" rely on that)

Anyway I'm afraid all of this is just crazy talking on meaningless numbers - but i guess you know that already :p

Regards,

"Helen Carolina Welton" by Jim Pick

Fri, 09 May 2008 09:43:49 -0700

Congrats!

"Helen Carolina Welton" by Filippo Diotalevi

Thu, 08 May 2008 06:24:27 -0700

Complimentoni!

"Helen Carolina Welton" by Martin Mainusch

Thu, 08 May 2008 06:07:52 -0700

Congratulations, David That's a very pleasant news! All the best for your family and You!