Summary
In this chapter, we have learned how to create matrices of any dimension in Python, how to access the matrices' elements, how to calculate basic linear algebra operations on matrices, and how to save and load matrices.
Working with NumPy matrices is a principal operation for any data analysis since vector operations are machine-optimized and thus are much faster than operations on Python lists—usually between 5 and 100 times faster. Backtesting any algorithmic strategy typically consists of processing enormous matrices, and then the speed difference can translate to hours or days of saved time.
In the next chapter, we introduce the second most important library for data analysis: Pandas, built upon NumPy. NumPy provides support for data manipulations based upon DataFrames (a DataFrame is the Python version of an Excel worksheet—that is, a two-dimensional data structure where each column has its own type).
