Learning SciPy

Dhanya Arumughom -

Python has a great ecosystem with some good libraries for scientific computing. This ecosystem consisting of various tools for scientific computing in Python is often called as the SciPy stack or NumPy stack. The stack consists of various Python libraries which assist scientists in solving their problems. Some of the important libraries in the stack are NumPy, SciPy, matplotlib, Pandas and SymPy. All these libraries together provide can provide a solution for a data analysis problem from data crunching to data visualization for predictive modeling.

I think you would have already guessed SciPy stands for “Scientific Python”. Don’t get confused with SciPy/NumPy stack and the SciPy and NumPy libraries with same name. The stack is named so because the NumPy and SciPy libraries are the foundation libraries for scientific computing in Python. You would need any of these fundamental libraries to work with other libraries. For example, Pandas and matplotlib are built on top of NumPy and Scikit-learn are built upon NumPy and SciPy.

What is in SciPy Library?

The SciPy library is a python library which provides a number of algorithms for math, science and engineering disciplines like numerical integration, various equation solvers, regression, optimization, linear algebra, signal and image processing and many other tasks. It is written in Python, Fortran, C and C++.

SciPy builds on the NumPy array object and this multidimensional array is the basic data structure used by SciPy. The NumPy library itself contains basic operations for scientific computing like some linear algebra functions, discrete Fourier transforms, etc. In fact, SciPy extends NumPy with further fully featured versions of the linear algebra and many other useful functions and numerical algorithms. Think of SciPy as an add on library to NumPy where NumPy provides the basic building blocks like vectors and matrices and provides operations on them. SciPy uses these building blocks to do more advanced things. So it could be said that most advanced features for computing are in SciPy rather than NumPy.

The SciPy library consists of many sub-packages which contribute to its computing capabilities and making it user friendly. Given below are few of the sub-packages used mostly.

  • cluster- For vector quantization/ Kmeans
  • fftpack – Discrete fourier transform algorithms
  • integrate – numerical integration routines
  • interpolate – integration routines
  • io – for data input and output
  • linalg- linear algebra algorithms
  • stats – Statistical functions
  • sparse – sparse matrix and related algorithms

Learn SciPy

NumPy and SciPy has led Python to make inroads into the data science world. Python which was conceived very long back around 1980s, a general purpose high level language, became a popular language for data analysis only very recently. Python was not suitable for statistics before the introduction of these packages. However, now it is equipped libraries for every aspect of data analysis. Combination of NumPy, SciPy, Pandas and matplotlib can be used for the entire data analysis process. Python with the SciPy stack has become one of the most preferred languages for data analysis and machine learning and has gained equal importance and fan followers as that of some of the other data languages widely used like the R programming language.

As mentioned before, NumPy and SciPy are the foundation packages for scientific computing using Python. So you need to start with these libraries for learning data analysis with Python.

Pre-requisites: Before you start to learn SciPy library you should have some basic knowledge of Python and the NumPy library.

Best Tutorials for SciPy

We have collected a list of some of the best beginner level tutorials on SciPy.

Best Courses For SciPy

However, if you already decided to take a deep dive, here are some of the best courses we found on SciPy.

For Beginner Level