>Inb4 slow and gay
I've been doing data science in Python for a while, and finally am working on a bigger project. What are the best resources for learning about Python package development and good Python practices (like the weird OOP stuff)?
I prefer books, but anything works...
>>60487281
Learn R instead
>>60487298
I know R but I would never use for something other than pure statistics or probability
>>60487281
>Python
>data science
hah, slow and gay
>>60487366
python is slow, the C backends are fast
>>60487679
this.
All the data science packages relay on Numpy which is just a glorified C code wrapper.
>>60487902
don't forget CUDA, hnngg at some of the gpu accelerated linear algebra libraries
>>60487281
Slow and gay, just like you.
>>60487281
Start with Raymond Hettinger (youtube lectures, more about python in general) and The Hitchhiker's Guide to Python.
I hope it will direct you where you want to go
>>60487679
The flaws about this idea are:
a) C<->Python can be a bottleneck
b) someone has to write and maintain those C backends
>>60488684
thank you lovely anon, this is exactly what I was looking for :)
please accept this attractive man as a token of my gratitude, any other advice will be equally well rewarded
>>60488735
>>>60487679 (You)
>a) C<->Python can be a bottleneck
barely, opencv slowdown from c++ to python is <1% according to the documentation
>b) someone has to write and maintain those C backends
...like any code ever?
>>60488804
>barely, opencv slowdown from c++ to python is <1% according to the documentation
That is, because as long you don't manipulate pixels yourself directly you are dealing with relative high level objects in opencv.
>>60488804
>...like any code ever?
People prefer working in decent languages.