R has been growing for years. It is more popular now then ever. Microsoft has bought into the ecco-system and I can use R with MS's sql server and also with Spark. People who had a bad experience from years ago and haven't gotten experience with all the changes we now have in R.
Python is a good choice but I think R is much better for my uses. Personally base 0 on Python with statistics is just wrong :) Python::Pandas has certainly gotten a lot of work to almost get equal parity with R.
Anything other than base 0 would be unacceptable in python; people would go crazy. Personally, this is the biggest reason I can't use R—my brain doesn't work with position-based indexes; I'm just hardwired to offset.
Just as you learned to start counting at 1 like the rest of the world, you can relearn it easy enough. I don't have any problems using Julia's 1 based arrays. I always thought 0 based in high level dynamic languages was silly.
What does dynamic vs static have anything to do with preferred indexing convention? The former is an execution time detail, the latter is how to index memory.
And that is 100% valid. So Python is a better choice. I just don't understand why the world isn't big enough for R and Python in the community especially since the leaders in R and Python get along so well.
> If you get a lot of value out of this package, do consider donating to HIBP since Troy Hunt does not put any limits on the API and it’s a tremendous service.
For those in the R ecosystem who want to do a broader analysis of what HIBP's encompasses, going through the API probably won't be very efficient other than to get a listing of the metadata.
I can't reach Troy's site right now, but I'm curious about the API itself: Does it handle Gmail "+" aliases? E.g., joe@gmail.com and joe+probablySpam@gmail.com route to the same mailbox. If I just search for joe, will breaches for joe+<whatever> be found?
Always nice to see these API extensions, but what exactly is the use case for this? According to haveibeenpwned.com there is a rate limit of 1 request per 1500 milliseconds per IP address, so this isn't efficient for analyzing breaches. It would be easier just to download a dump of the leaked data, clean it, and import it into R.
It's great that they're doing this. We had kicked around the idea of signing up all of our customers for HIBP, but decided against it for obvious reasons. What we would really like to do is to be able to notify our custoemrs when their email addresses show up in breaches. Besides being helpful to our customers (who will mostly not be aware of useful services like HIBP), it would potentially help reduce fraud on our sites.
Is anyone aware of a way to get access to sanitized dumps that we can compare to our customer DB internally? It's unlikely we'd get approval to go out and get the dumps and analyze them ourselves, but if there was a reliable source like HIBP, but for bulk comparison, we could bring a lot of value to the business and to our customers.
15 comments
[ 2.2 ms ] story [ 46.5 ms ] threadHello matlab, hello fortran! Yes scientists still use both of these.
R has been growing for years. It is more popular now then ever. Microsoft has bought into the ecco-system and I can use R with MS's sql server and also with Spark. People who had a bad experience from years ago and haven't gotten experience with all the changes we now have in R.
Python is a good choice but I think R is much better for my uses. Personally base 0 on Python with statistics is just wrong :) Python::Pandas has certainly gotten a lot of work to almost get equal parity with R.
I believe there is a rate limit: one request per 1.5 seconds from a given IP address https://haveibeenpwned.com/API/v2#RateLimiting
For those in the R ecosystem who want to do a broader analysis of what HIBP's encompasses, going through the API probably won't be very efficient other than to get a listing of the metadata.
https://haveibeenpwned.uservoice.com/forums/275398-general/s...
Is anyone aware of a way to get access to sanitized dumps that we can compare to our customer DB internally? It's unlikely we'd get approval to go out and get the dumps and analyze them ourselves, but if there was a reliable source like HIBP, but for bulk comparison, we could bring a lot of value to the business and to our customers.