What Python can do in data analysis, but MATLAB cannot?

I am familiar with MATLAB and often listen to Python recommended by others. It is time-consuming to learn a new language. What Python can do for data analysis, but MATLAB cannot? I am familiar with MATLAB and often listen to Python recommended by others. It is time-consuming to learn a new language.
What Python can do for data analysis, but MATLAB cannot? Reply: I personally have Python and MATLAB. I mainly use C ++ to implement Intelligent Algorithms for data analysis. As far as I know, these languages are competent for all data analysis tasks. The difference is nothing more than the difficulty level. What you need to do is to use your most skilled language as much as possible, focus on theory, algorithm, and actual modeling.

Yes. Basically, intelligent algorithms can be implemented in any language. intelligent algorithms are only related to relevant theories, algorithm design, and actual modeling, and are irrelevant to specific languages. However, the toolkit provided by different languages is different, which makes it difficult to implement algorithms. As far as I know, MATLAB and Python provide almost as many data packets in the field of data analysis. So I suggest you continue to enhance your own familiar MATLAB. If you are interested in time, you can consider learning Python.

If you really want to engage in the big field of artificial intelligence and data mining, I suggest you increase your energy from the most basic programming language to the basic principle, algorithm design, and actual modeling. This may be more beneficial. After learning these three types of libraries, you can easily complete data analysis even if you use a poor language such as C ++. In turn, if you are not familiar with this and will only use the toolkit to run a result, you not only do not understand the method, but what's more terrible is that you may not build models based on specific problems; such a result can only be a low-entry level. This is my opinion.

PS: after reading the answer, I think my answer is limited to the Core Algorithm of data analysis. If you need to complete a data analysis system from start to end, it must be as @ Lu Qiang said, python is better. However, I still want to emphasize that the core algorithm is also the top priority in the complete data analysis system, so I strongly recommend that you increase your understanding of the method itself. There are many examples. Depending on your specific application, python is a universal language, while matlab is a matrix operation, which is inherently inadequate in terms of general language functions, such as string processing mentioned above, matlab can do but it is certainly not as comfortable as python. If you want to perform Chinese Word Segmentation and natural language processing, it may be more troublesome.

To give an example, I can also give one. I need:
1. Deploy a distributed crawler on aliyun to regularly capture data through the Internet every day, and then clean and store the data. If you have PostgreSQL and mongodb.
2. Because SQL data may change the data structure, I also need to Use orm to manage version changes to the database table structure.
3. After finishing the data, I need to use a certain model for analysis and then dynamically generate some charts.
4. I want to create a dynamic web page to display relevant data and analysis charts.

The above work (with my understanding of matlab) matlab can only complete step 1, while python can all be done, and install a pirated version on aliyun (of course I can't afford a genuine version) linux-based matlab is also quite painful to think about. Of course, if your requirement is only step 1, you can use matlab.

Certificate ---------------------------------------------------------------------------------------------------------------------------------------
Ps: I don't know how many data analysis tasks are actually performed using matlab in the production environment, but I know that python is actually used a lot. In the production environment, you must consider the cost of matlab. Therefore, if you want to engage in data analysis rather than scientific research, you are advised to learn python, a free language, although python may not have more than a matlab database in terms of numeric data analysis, there are more things that python can do, you can use only one language from data capture, cleansing, sorting, analysis, and presentation, and python is really easy to learn.

There is no meaning for matlab here. I have used matlab since 98 years and helped me complete many projects, so I still have a good impression on it, data processing can greatly improve development efficiency. When I leave school, I use less and less products. Piracy is too inconvenient, and I have a very low desire to install such an unfrequently used giant on my own machine. In python, any linux release comes with its own version, and you can simply install a 100 or 200 MB tool. The MATLAB operator string is uncomfortable, although it seems to be capable.

In this way, it is uncomfortable to do the work of bioinformatics which requires frequent string operations.

I basically didn't use MATLAB as a new mail around me, although it has this package. It has been nearly a year since the last question. Based on your own practice, you can answer your questions.
1. Data cleansing, data preprocessing, and Python pandas explosion MATLAB.
2. In machine learning, Python scikit is too easy to use. MATLAB has few algorithms in this aspect, and hyperparameters need to be optimized through a loop, while scikit-learn has one or two sentences.
3. Deep Learning Methods: Python has theano, keras, lasagne, and so on. There are almost no libraries in this aspect of MATLAB. One of them is found and maintenance is stopped, we also recommend that you use the Python library...
4. The most obvious advantage of MATLAB over Python is its IDE, which is convenient for debugging and comprehensive in documentation. If someone else processes the data for you, that would be a lot better. Data crawling by yourself, or adding standard or third-party libraries to the conversion format Python. A lot of MATLAB cannot do, But Python is easier in most cases, such as reading sqlite databases and writing GUI programs, when operating system interaction is required, IronPython interacts with COM, or you need to change for to parfor in MATLAB, and then find that the company didn't buy parallel toolbox ......

In a word: Python has the import algorithm, which makes MATLAB a little difficult ......

Well, apart from the convenience of reading databases and the need for parallel computing, other items are not necessary in data science ...... Both MATLAB and Python are completely Turing-complete, and Python can do the same with MATLAB! In some ways, the two languages are not as good as the other ......