[data analysis and display]

conda

A tool , For package management and environment management
Package management and pip similar
Environmental management allows users to use different versions python, And can switch flexibly

The dimensions of data

The organization of a set of data
Lists and arrays
The ordered structure of a set of data
difference ： list ： Data types can be different
Array ： Same data type
Two dimensional data ： Multiple one bit data constitute , Is a combination of one bit of data
Multidimensional data ： One dimensional and two-dimensional data are expanded in new dimensions to form

High dimensional data ： Use the most basic binary relationship to show the complex structure between data

numpy——scipy,pandas Data processing or science · Basis of calculation library

1, One Powerful N Dimensional array object ndarray
2, Broadcast function
3, Integrate c/c++ Code tools
4, linear algebra , The Fourier transform , Random number generation and other functions
introduce ：
import numpy as np
ndarray:N Dimensional array object
The array object can remove the loop required by the operation between elements , Make a one-dimensional vector more like a single data
Set special array objects , optimized , It can improve the computing speed of this kind of application
General , The type of all data in a dimension is often the same . Array objects use the same data type , It helps to save computing and storage space
ndarray Two parts make up ：
1, The actual data
2, Metadata that describes the data （ Data dimension , Data type, etc ）
Axis ： Save the dimension of the data , Rank ： Number of shafts
ndarray attribute ：.ndim Rank , Number of dimensions ·
.shape Object size
.size Element number
.dtype Element type
.itemsize Element size

establish ndarray

（1）x = np.array(list/tuple)
When np.array（） Don't specify dtype when ,numpy Will associate a based on the data dtype type
1, Create from list type
2, Create from tuple type
3, Create from list and tuple mixed types
（2） Function creation
arange,ones,zeros etc.
eg：np.arange(10)
np.ones((3,6))
(3) Other functions create

ndarray Array transformation

Dimensional transformation
.reshape(shape) Do not change array elements , Return to one shape Array of shapes the original array remains unchanged
.resize(shape) Modify the original array
.swapaxes(ax1,ax2) Will array n Two of the three dimensions are exchanged
.flatten() Dimension the array , The original array remains the same
ndarray Array type transformation
new_a = a.astype(new_type)

Index and slice of arrays

A one-dimensional ： And python The list of is similar to

Index of multidimensional array

Slices of multidimensional arrays

The operation between an array and a scalar

The operation between an array and a scalar acts on every element of the array
numpy One variable function


numpy Dual function