I used this at one point when I needed a large synthetic "customer" dataset, and found the sample size to be somewhat limited. I had to munge first/last names with added random characters. But looking at it now, it appears that's improved. For example, there are about 3000x500 first/last name combinations, for a total of about 1.5M possible full names (for en_US). However, if you need, say, as many unique last names as possible, 500 might still be too few. Still, a nice project.
I always liked this library. Wrote a tutorial around the concept of it on websesign tuts a few years back, though I'm not sure how dated it is in relation to the API now (http://webdesign.tutsplus.com/tutorials/filler-content-tools...). Anyway, it deals with one way the library could be used in the browser, in combination with other filler content services, to aid "designing" in the browser.
Could be used for both. I've used it for generating test data in unit/integration tests and also for obfuscation where production-dumps are needed.
It's useful for having in development as well, when you need to have X number of forms and X number of users but you don't want to have to manually create that every time you re-create the environment.
It can help a lot when writing UI/Selenium tests with mocked/fake backend (as long as the application data model is well defined).
i also wrote a similar library around es6 generators that defines application data structures as plain JS objects and generates data objects out of it using fake random data
https://github.com/dmitriiabramov/sharkhorse
Faker.js is pretty great. I wrote a simple package to generate fake data for MySQL tables. It's rather basic, but it's been useful to me. It probably wouldn't be hard to stick a DB abstraction layer in front to use with other backends.
Good idea. Can anyone comment on this vs traditional QuickCheck like tools? We use property-based testing as our default. It seems like Faker is a less composable version.
[0] https://github.com/RaedsLab/Tunisian_Name_Generator