Data Analysis by Python

Plot of Panda (matplotlib)

Panda is one of the packages of Python. We can use data type of Multi-Variable Analysis in Python when we use Panda.

"Plot" of Panda is the function for Graphical Analysis . "Plot" is made with Panda and matplotlib.

Common Code

This code set is needed before the code starting "df".

import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline # To draw graphs in Jupyter
df= pd.read_csv("Data.csv")

Short Passes to the Graphs

Graph to visualize many variables
>>>Line graph for all variables
>>>Line graph in different space
>>>Scatter plot for one variable and others

Histgram
>>>2-dimension Histgram

Memo
>>>Size of Graph

Graph to visualize many variables

Line graph for all variables

df.plot()

Very simple code! So it is useful as the first step of the data analysis.

But in this example, we do not know details of variables except X1. So we separate graphs for each varaiable.

df.plot(subplots=True)

df.plot(subplots=True, sharey=True)# Ranges of Y axis is same

Line graph in different space

If the variable of X axis is set, x axis is not same space.

If one varaiable is set as X axis, line graphs are made for other variables.

df.plot(x = 'X1',marker="o")

df.plot(x = 'X1', subplots=True, marker="o")

df.plot(x = 'X1',y ='Y1',marker="o")

Scatter plot for one variable and others

df =df.sort_values('Y1') # Sorting for X variable
df.plot(x = 'Y1', subplots=True, marker="o", linestyle='None')

Histgram

This page is only the 2 dimenion histgaram. General histgram is written in seaborn page because combination with seaborn is useful.

2-dimension Histgram

We can make 2-dimension Histgram by matplotlib not by Plot of Panda.

plt.hist2d(x=df.X1,y=df.Y1)

Memo

Adjust Size of Graph

Left graph is default size. Right is the adjusted size.

This code is same for the graphs starting the code "df.".

df.plot(subplots=True,figsize=(3,3))

Plot of Panda and seaborn

Plot of Panda is good at to use left type data. For the right type data, we need to write many codes.

seaborn is good at to use right type data.

I use both Plot of Panda and seaborn in the same code set.