Skip to main content
SciPy

SciPy

Computer programWikipedia

Search complete. 39 mentions across 14 episodes found for "SciPy".

Nov 22, 2026

David CloughHOST
76:28
So that gives you that.
David CloughHOST
76:30
And then we come along and you end up getting like specific manuals for NumPy, Another module we use frequently later on is called SciPy.
David CloughHOST
76:45
And then there's a user's guide to Matplotlib.
David CloughHOST
76:50
So you can do that.
Philipp HennigGUEST
27:56
And you see these kind of work precision diagrams for many different solvers in many different languages actually across language ecosystems.
Philipp HennigGUEST
28:04
Here at the top in gray is MATLAB, in green is SciPy, in blue is old Fortran code by Ernst Heyer.
Philipp HennigGUEST
28:12
and professional C++ code.
Philipp HennigGUEST
28:14
And then down here are the super optimized Julia libraries.
Harrison KinsleyHOST
9:18
And then how to resample.
Harrison KinsleyHOST
9:20
I'm pretty sure we can use SciPy for resampling information.
Harrison KinsleyHOST
9:25
But anyways, that's the question.
Harrison KinsleyHOST
9:27
And now how do we get the data actually just in Python? So to do that, it was actually super simple.

6 MINS LATER

Harrison KinsleyHOST
15:54
But like I said, I'm pretty sure gamma and beta is going to give us the information that we need, you know, for control.
Harrison KinsleyHOST
16:01
Like if we're thinking left, thinking right, I think we're going to need somewhere in these range.
Harrison KinsleyHOST
16:07
If I forgot to say it too, I'm pretty sure I can use SciPy for the actual resampling, but it really is going to depend on if this is seamless or not.
Harrison KinsleyHOST
16:18
I think it's bursting.
Harrison KinsleyHOST
4:39
um basically the crux is greater than two so if you want 60 hertz you need a raw sample sampling that is more than 120 hertz so that makes sense because also on the open bzi gui if you're you know like i'm i'm using 125 hertz it goes to 60 okay so i'll take that as expert validation so So moving on, now that we know those things, what is our objective? Our objective is at least to start, I want to move something on the screen, and as you've already seen, I'm starting to get there, but how do we do it? So we need something, and I want to use deep learning, so we need something to pass through this neural network model that is useful information.
Harrison KinsleyHOST
5:25
So I did start with PyOpenBCI and I was doing my own signal processing with SciPy.
Harrison KinsleyHOST
5:31
This is fairly complex.
Harrison KinsleyHOST
5:33
It's not that hard, but you still need to account for any fluctuations in the actual sampling rate.
speaker_0NARRATOR
0:02
Data scientists often praise NumPy for its power.
speaker_0NARRATOR
0:06
SciPy looks at NumPy and thinks, adorable.
speaker_0NARRATOR
0:09
SciPy is not here to perform basic calculations.
speaker_0NARRATOR
0:12
It integrates, optimizes, transforms, and solves.
speaker_0NARRATOR
0:17
This is a library for those who have mastered simplicity and are prepared to embrace complexity.
speaker_0NARRATOR
0:23
Let us explore SciPy, where mathematics meets precision engineering.
speaker_0NARRATOR
0:28
Shall we begin? Install SciPy.
speaker_0NARRATOR
0:33
SciPy is no ordinary library.
Paul AllenGUEST
64:27
So Travis Oliphant, who just moved to Utah after 25 years in Austin, you guys should interview Travis.
Paul AllenGUEST
64:33
He invented NumPy and SciPy.
Paul AllenGUEST
64:35
He created Anaconda, which is a $1 or $2 billion company to distribute Python into enterprise.
Paul AllenGUEST
64:41
Travis is one of the top three or four people in the world in the Python ecosystem.
Paul AllenGUEST
64:46
He's the reason Python was the language for AI.
Paul AllenGUEST
64:50
Yeah.
Paul AllenGUEST
64:50
because he built NumPy and SciPy.
Paul AllenGUEST
64:54
So Travis is an investor and we hired some of his AI experts to build custom ML models back in 2020.
RJ HanekeHOST
56:07
So why is this an important problem if we can just solve them with tools? I mean, maybe you can't this one specifically, but.
JonasGUEST
56:15
Yeah, so we can't, even when we do run these as agent tasks and we provide a bunch of Python libraries for scientific purposes like SciPy and a bunch of image manipulation libraries like OpenCV, the accuracy is significantly better than the VLMs, but it's still under 60%.
JonasGUEST
56:33
And so, yeah, like the...
JonasGUEST
56:35
It's still nowhere near usable.
Harrison KinsleyHOST
1:23
Now let's say you had a lot more plots or something, and you really wanted to make this smooth, right? You don't want this jagged edge, the jagged edge.
Harrison KinsleyHOST
1:32
There's a very quick way about doing this, and the quick way that I know of anyways, comes from a download of SciPy.
Harrison KinsleyHOST
1:40
So what you'll want to do is come over to your favorite search engine, which is probably Google, and you'll want to do a SciPy download.
Harrison KinsleyHOST
1:50
Let me move this over so you can see it.
Harrison KinsleyHOST
1:54
Just scipy download.
Harrison KinsleyHOST
1:56
And then the first option for you will probably be the sourceforge.net. Click on that and then you'll click on this download for your computer.
Harrison KinsleyHOST
2:08
Now, if you're on Mac OS or you don't have Python 2.7 or whatever, first of all, get Python 2.7. And then Windows 32, depending on whatever your operating system is, you might want to pick something else.
Harrison KinsleyHOST
2:22
But just make sure you get SciPy.
Grant SandersonNARRATOR
22:34
And in this case, it runs it for multiple different iterations, tries to find an average, and it looks like, on this computer at least, it averages at 4.87 seconds.
Grant SandersonNARRATOR
22:43
By contrast, if I use a different function from the SciPy library called fftConvolve, which is the same thing, just implemented differently, that only takes 4.3 milliseconds on average, so three orders of magnitude improvement.
Grant SandersonNARRATOR
22:57
And again, even though it flies under a different name, it's giving the same output that the other convolve function does, it's just doing something to go about it in a cleverer way.
Grant SandersonNARRATOR
23:08
Remember how with the probability example, I said another way you could think about the convolution was to create this table of all the pairwise products, and then add up those pairwise products along the diagonals.
Nathan GoldbaumGUEST
12:51
It's not.
Nathan GoldbaumGUEST
12:52
OpenBLAST, at one point, when NumPy depended on Atlas, SciPy has a lot of Fortran code, but NumPy has never had any Fortran code.
Nathan GoldbaumGUEST
13:02
It's all C.
Nathan GoldbaumGUEST
13:03
It talks to CPython directly over the CPython C API.

26 MINS LATER

Nathan GoldbaumGUEST
38:44
Yeah, it's interesting that way, yeah.
Nathan GoldbaumGUEST
38:46
And often the, like, social part of it, I think, is the most important.
Nathan GoldbaumGUEST
38:50
And that's why conferences like the Scientific Python Conference, I gave that talk you're talking about at SciPy 2024 in Tacoma.
Nathan GoldbaumGUEST
38:55
And, like...

4 more episodes mention SciPy.

Create an account to see the whole feed, search across every transcript, and follow the entities you care about.

We value your privacy

We use cookies to understand how you use our platform and to improve your experience. Click “Accept All” to consent, or “Decline non-essential” to opt out of non-essential cookies. Read our Privacy Policy.