Support different decimal/thousands separators

[schnerd]

Is your feature request related to a problem? Please describe.
There are numerous international datasets that Kepler cannot process because they use different decimal separators other than a period, such as this Buenos Aires Taxi Dataset, which uses a comma instead of a period.

In addition to commas instead of decimals, there are a few other formatting types that may be worth supporting: https://en.wikipedia.org/wiki/Decimal_separator#Examples_of_use

Describe the solution you'd like
I think in an ideal world Kepler would automatically detect the number formatting type, in a similar way that it detects the data type. However, I think letter people enter custom thousands/decimal separators is another option (numeral.js takes this approach).

I think this type of input parsing problem isn't too dissimilar from supporting CSVs with different delimiters (#196). It seems like the upload dialog could maybe have a couple more configuration options after upload–including changing the column delimiter, thousands/decimal separators, and possibly re-adjusting datatypes (#191).

Describe alternatives you've considered
You can transform this data use excel or another python script, but it would be awesome to support these datasets out of the box.

[macrigiuseppe]

linked to: uber-web/type-analyzer#7

[rusosnith]

What happened to this? It's so US centric to have a tool only compatible with the units/delimiters used in one country.
This could be implemented so easily!