5 ms·

Faker.js - generate massive amounts of fake data in the browser and Node.js

107 points by daGrevis 14y ago

DoubleCluster 14y ago
What's the use case for this? To annoy spammers?
- checktiput 14y ago
  Debug I suppose.
- dugmartin 14y ago
  I built something like this for an internal CMS framework and it really speed up prototyping pages - especially large list pages.
- Swizec 14y ago
  Coming up with random words and examples when writing tests is surprisingly tedious.
schumacher 14y ago
Cool. I recently build a similar thing for Magento to provide demo data and for anonymization purposes. Some parts of faker.js needs sure refactorings ...
- agotterer 14y ago
  I've been looking for something that can generate fake cart data. Is what you built open source?
revorad 14y ago
This is fantastic. I was actually looking for something like this to test some APIs with dummy data.
- paavo 14y ago
  The original such library was Perl's Data::Faker which was ported to Ruby as Faker gem (and later refactored to faster and more comprehensive Ffaker gem).
qwerty69 14y ago
Very nice library. It would be nice to add the ability to generate random values based on a distribution, e.g. the values are random but correspond to a normal distribution with a given mean and standard deviation. This would be very useful for testing statistical libraries and applications.
- SammoJ 14y ago
  Which attributes would you sample from a normal distribution here? I don't see any numerical attributes where this would make sense. One could add weight, height, age etc. and sample from the relevant geographical/gender distributions?
  - qwerty69 14y ago
    An example would be the processing of automatically acquired measurement data. Such data usually shows some kind of distribution (binominal, normal, geometric, etc...) that needs to be taken into account when validating such a tool.
DanBC 14y ago
This is cool. I'm gently disturbed that it generates valid email addresses. You pick a random name, follow it with an @, then add a random word, and then add a TLD. Obviously you're not responsible for how people use the tool. Best practice would be to use the domain example.com
- paavo 14y ago
  In most use cases I can think of the generated data is used only in a closed (testing) environment and won't be visible to web crawlers and spambots. Many developers like me use these libraries for prototyping different UI views for their apps. Unfortunately, spammers do have way more advanced technologies to try and guess working addresses. Libraries like this generate mostly very improbable addresses: you would get way better results just by taking lists of most popular first names and family names and concatenating them with some well-known web mail providers.
  - mindcrime 14y ago
    In most use cases I can think of the generated data is used only in a closed (testing) environment and won't be visible to web crawlers and spambots. Many developers like me use these libraries for prototyping different UI views for their apps. The bigger problem, in my experience, is when you accidentally spam people yourself, while testing some function that somehow generates an outgoing email. I think almost everybody that's ever worked on code that sends emails has - at one time or another - ran a job that somehow wound up sending 6,712,374 "test" messages to every user in the production database. :-) Dummy addresses are better, but it's better still if they use @example.com so you know there's no way they'll actually get delivered to anybody.
- bpicolo 14y ago
  If someone really wants to guess random addresses it isn't hard for them to write it themselves.
tonyblundell 14y ago
Great, thanks!
webfuel 14y ago
Thanks! I won't be using http://www.fakenamegenerator.com/ http://www.fakenamegenerator.com/ as much now.
fruchtose 14y ago
Great project! I would, however, recommend an API with more namespaces. Functions like Address.ukCounty don't sit well with me; such an API would be a mess if you added support for countries other than the US and the UK. You could (and probably create namespaces for each country code, e.g.: Address |-> US |-> state |-> UK |-> state An alternative would be for country-specific generators to take a country code as an argument. I am not an experienced API designer, though, so take my comments with a heaping of salt.
justwrote 14y ago
Shameless plug: I also created a similar library for Scala some time ago https://github.com/justwrote/scala-faker https://github.com/justwrote/scala-faker Need to update to 2.10 soon :)
robbyishere 14y ago
anjuna.js
hawkw 14y ago
So many Node.js libraries....
mindcrime 14y ago
Obligatory "I wrote something similar" post. But my tool is written in Java, doesn't run in the browser, but does use @example.com for the email addresses. https://github.com/mindcrime/dummydatagenerator https://github.com/mindcrime/dummydatagenerator

107 points by daGrevis 14y ago

DoubleCluster 14y ago

What's the use case for this? To annoy spammers?

checktiput 14y ago
Debug I suppose.
dugmartin 14y ago
I built something like this for an internal CMS framework and it really speed up prototyping pages - especially large list pages.
Swizec 14y ago
Coming up with random words and examples when writing tests is surprisingly tedious.

schumacher 14y ago

Cool. I recently build a similar thing for Magento to provide demo data and for anonymization purposes. Some parts of faker.js needs sure refactorings ...

agotterer 14y ago
I've been looking for something that can generate fake cart data. Is what you built open source?

revorad 14y ago

This is fantastic. I was actually looking for something like this to test some APIs with dummy data.

paavo 14y ago
The original such library was Perl's Data::Faker which was ported to Ruby as Faker gem (and later refactored to faster and more comprehensive Ffaker gem).

qwerty69 14y ago

Very nice library. It would be nice to add the ability to generate random values based on a distribution, e.g. the values are random but correspond to a normal distribution with a given mean and standard deviation. This would be very useful for testing statistical libraries and applications.

SammoJ 14y ago
Which attributes would you sample from a normal distribution here? I don't see any numerical attributes where this would make sense. One could add weight, height, age etc. and sample from the relevant geographical/gender distributions?
- qwerty69 14y ago
  An example would be the processing of automatically acquired measurement data. Such data usually shows some kind of distribution (binominal, normal, geometric, etc...) that needs to be taken into account when validating such a tool.

DanBC 14y ago

This is cool. I'm gently disturbed that it generates valid email addresses. You pick a random name, follow it with an @, then add a random word, and then add a TLD. Obviously you're not responsible for how people use the tool. Best practice would be to use the domain example.com

paavo 14y ago
In most use cases I can think of the generated data is used only in a closed (testing) environment and won't be visible to web crawlers and spambots. Many developers like me use these libraries for prototyping different UI views for their apps. Unfortunately, spammers do have way more advanced technologies to try and guess working addresses. Libraries like this generate mostly very improbable addresses: you would get way better results just by taking lists of most popular first names and family names and concatenating them with some well-known web mail providers.
- mindcrime 14y ago
  In most use cases I can think of the generated data is used only in a closed (testing) environment and won't be visible to web crawlers and spambots. Many developers like me use these libraries for prototyping different UI views for their apps. The bigger problem, in my experience, is when you accidentally spam people yourself, while testing some function that somehow generates an outgoing email. I think almost everybody that's ever worked on code that sends emails has - at one time or another - ran a job that somehow wound up sending 6,712,374 "test" messages to every user in the production database. :-) Dummy addresses are better, but it's better still if they use @example.com so you know there's no way they'll actually get delivered to anybody.
bpicolo 14y ago
If someone really wants to guess random addresses it isn't hard for them to write it themselves.

tonyblundell 14y ago

Great, thanks!

webfuel 14y ago

Thanks! I won't be using http://www.fakenamegenerator.com/ http://www.fakenamegenerator.com/ as much now.

fruchtose 14y ago

Great project! I would, however, recommend an API with more namespaces. Functions like Address.ukCounty don't sit well with me; such an API would be a mess if you added support for countries other than the US and the UK. You could (and probably create namespaces for each country code, e.g.: Address |-> US |-> state |-> UK |-> state An alternative would be for country-specific generators to take a country code as an argument. I am not an experienced API designer, though, so take my comments with a heaping of salt.

justwrote 14y ago

Shameless plug: I also created a similar library for Scala some time ago https://github.com/justwrote/scala-faker https://github.com/justwrote/scala-faker Need to update to 2.10 soon :)

robbyishere 14y ago

anjuna.js

hawkw 14y ago

So many Node.js libraries....

mindcrime 14y ago

Obligatory "I wrote something similar" post. But my tool is written in Java, doesn't run in the browser, but does use @example.com for the email addresses. https://github.com/mindcrime/dummydatagenerator https://github.com/mindcrime/dummydatagenerator