[wyldfire]

> Tourism graph is based on Google searches for services used by tourists relative to the rest of the year.

That's a pretty clever way to go about it. I hope/suppose that the sampling bias isn't correlated with time-of-year somehow.

[clementangerine]

It seems pretty off, though. For example, the page for Iceland says that winter is the peak time, but it's actually pretty dead in the winter with the peak in the summer. Seems like people start planning their summer trips around December (plane tickets for Christmas?). Other places seem shifted up a month or two - the page for Paris has the peak season around May-June, but it's actually more like June-August.

OP, if you want to improve this, could you get some actual tourist data and experiment a bit with which other metrics (various Google searches - something like taxi services might be more accurate to estimate how many people are actually traveling in the country at the time, hotel prices, plane ticket prices, etc) best predict the tourist numbers?

[bawana]

It's actually inaccurate. People search BEFORE they go. In fact, I can tell you the busiest tourism month in Greece is August. I've been there dozens of times and June is actually a pleasant (tho cool-ish) time to visit - no crowds. And the Greeks are still working (trying to make money to spend on their August vacations - when everything shuts down). Perhaps Airline charters would be a more accurate reflection. Altho I do not know where to source that data.