Python for decision makers and business leaders Transcripts
Chapter: Python vs.
Lecture: Python vs. MATLAB
0:00 Python vs Matlab.
0:01 This is mostly applicable if you're doing data science
0:05 scientific computing, data visualization, things like that.
0:09 So, let's put them side by side. Is Matlab open source?
0:13 No way, no it is not. And Python, yes of course.
0:17 Is it compiled? No, these are not compiled.
0:19 Is it owned by a company?
0:20 Yes, Matlab is owned by a commercial identity
0:22 and Python is not. Do they have base-class libraries?
0:25 Yeah, Matlab actually has super good base-class libraries.
0:28 That's one of it's features, it gives you all these
0:31 libraries to do scientific computing and mathematical stuff.
0:35 Web capabilities? No.
0:37 It's not for that, and Python, it's very strong.
0:39 Database capabilities? Not really.
0:41 Not in a general sense.
0:43 Though, yeah, Python's very strong.
0:44 Mobile app capabilities? No.
0:47 Python is poor, but I believe Matlab is even worse.
0:50 Desktop capabilities? No, not really.
0:53 I know you can certainly automate Matlab
0:57 but when you distribute it you have to buy licenses
1:00 to Matlab and other funkiness
1:01 so, I'm going to say no there.
1:04 And again, Python does not win this super well.
1:06 They have Qt and a few other things
1:08 WX python, for example
1:10 but it wins in this competition here.
1:12 Stack Overflow rank?
1:13 Matlab does not appear on the list.
1:16 I'm sure there are questions about Matlab
1:18 on Stack Overflow but far as the analytics
1:21 I'm going to put negative infinity here
1:23 vs. the number 1 position.
1:24 That's a big difference.
1:26 TOB rank? Matlab doesn't appear there either.
1:28 Even though there is a programming language there.
1:30 Price? The price for Matlab is non-trivial.
1:33 This is $860 per user, per year, in perpetuity.
1:38 It goes on and on.
1:39 So, you can actually buy a license for something
1:42 like three times that, and it will give you a
1:44 permanent license forever, all of those are infinitely more
1:47 than free and this one is a huge, huge
1:50 importance distinction, is Matlab a general purpose
1:53 programming language?
1:54 No, it's not a general programming language you can't build
1:57 arbitrary things can you go build YouTube with it?
2:00 No! Would you ever try to build Microsoft Word with Matlab
2:04 No, that would be insane you can not do it.
2:06 It's not built for that, why is that important?
2:09 Because if you start doing you're analysis in your work
2:11 in something that is not a general purpose language
2:14 Matlab or Julia whatever, stay focused on Matlab, we
2:18 build something in Matlab and we get it working
2:20 and it's like ah this is really cool, couldn't we turn this
2:22 into a great website or a cool product or, integrate this
2:25 into our other thing? The answer very likely is no you can't.
2:30 Because Matlab doesn't do that broader thing where as with
2:33 Python, if you start building up your Jupiter notebooks
2:36 and get some analysis going and then it's super easy to
2:39 productionize that and put that into a service or package it
2:42 up into an application, because you know what?
2:44 Python does those things but this is a big
2:46 big point about this general purpose programming language
2:50 and, I think it's really important in the consideration
2:52 you're making and then we have the scientific
2:54 compunioun level, I'm tempted to put Matlab under very
2:57 strong and Python also under very strong.
3:00 But why didn't I give Matlab as high of a score?
3:04 If I had unlimited money, I'd believe they would be on par.
3:08 But in order to do the extra stuff outside of the box
3:12 of Matlab. What you do is you go and buy these tool boxes, you want
3:16 to do wavelit decomposition?
3:17 Cool that's a $1000 dollars per user, per year to add
3:21 that feature to Matlab, where as with Python
3:24 that's just another free thing you get
3:26 off pipeei if you had unlimited money, I think their very
3:29 much on par with each other, but because they go outside
3:31 the box in Matlab you have to buy additional things
3:35 and those are extremely non-trivial prices, I think that
3:37 kind of takes a little bit away from the broad capabilities
3:42 of Matlab. So, at least at the base level let's say Python and Winscp
3:45 my interpretation is that it wins because you can go get all
3:48 these other libraries like astropy and so on.