[Python] Introduction to graph creation using coronavirus data [For beginners]

Introduction

I will explain the series of steps from reading a CSV file based on WHO public data to graphing it.

The site I referred to is here.

The above site is explained on the assumption that you have a basic knowledge of Python and matplotlib.

In this article, I will summarize the points that I did not understand. Supplementary explanations are written so that even beginners can understand.

Premise

• Environment setup such as Python and Jupyter Notebook has been completed.
• The CSV file to be read is published on the github link of Reference site, and this time I am using it.
• This content is for those who are about to learn machine learning. (For those who already understand machine learning, it may be a little unsatisfactory.)

environment

• macOS High Sierra 10.13.6
• Python 3.7.6
• matplotlib 3.1.3

goal

Our goal is to:

• Read CSV
• You can draw graphs with matplotlib

Preparation

Data reading

First, load the CSV.

import pandas as pd

#Data capture
df = pd.read_csv("COVID-19.csv", index_col='Date', parse_dates=['Date'])

I'm importing a library called pandas and reading a CSV file. pandas is a library that provides easy-to-use data structures and data analysis tools such as CSV loading.

For the basic contents of Japanese, you can understand what you can do by looking at here.

Supplement

The arguments for pandas.read_csv () are:

• 1st argument: CSV file to be read
• 2nd argument: index_col… Specify index column
• Third argument: parse_dates… Convert Date column to DateTime type

There are many other parameters available, so it's confusing at first, but I think it's okay if you understand them one by one.

Data display

Import matplotlib

python

import matplotlib.dates as mdates

It refers to a library called matplotlib. matplotlib is a library that allows you to create static, animated, and interactive visualizations in just a few lines.

Among them, this time, we will use matplotlib.dates to handle time series data with the horizontal axis as the date.

You can add an alias by adding ʻasto the import statement. You can call it by writingmdates instead of matplotlib.dates` each time.

Get color pattern

python

#Get color pattern
cmap = plt.get_cmap("tab10")

This time, each line needs to be color-coded in order to display multiple data on one graph. It is difficult to think about colors one by one, so it is a library that has an appropriate color pattern.

It is an image that switches the cmap to be read according to the number of lines to be displayed.

Use a lot of colors with matplotlib

Definition of scale

python

#Define the scale to set
locator = mdates.AutoDateLocator()
formatter = mdates.ConciseDateFormatter(locator)

When creating a graph, there are various decisions such as the unit of scale. The above process can be briefly explained as follows.

• AutoDateLocator () allocates axis scales nicely.
• Use ConciseDateFormatter (locator) to reflect a concise time series label on the locator.

Reflect the defined scale information on the graph

python

ax = plt.gca()

With plt.gca (), the current Axes object is acquired and set as the object of ʻax`.

Supplement What is an Axes object?

It seems that there are many people who have doubts about Axes objects and Figure objects.

The image is as follows.

Source: Official document

If you want to understand the above in detail, the following articles may be helpful.

[Basic knowledge of matplotlib that I wanted to know early, or the story of an artist who can adjust the appearance](https://qiita.com/skotaro/items/08dc0b8c5704c94eafb9#figure-axes-axis%E3%81%AF%E9 % 9A% 8E% E5% B1% A4% E6% A7% 8B% E9% 80% A0% E3% 81% AB% E3% 81% AA% E3% 81% A3% E3% 81% A6% E3% 81 % 84% E3% 82% 8B)

Drawing a graph

#Reflect the defined axis scale on the graph
ax.xaxis.set_major_locator(locator)
ax.xaxis.set_major_formatter(formatter)

#drawing
ax.plot(df.index, df['Global Confirmed'], "s-", color=cmap(3))
ax.plot(df.index, df['China Confirmed'], "o-", color=cmap(2))
ax.plot(df.index, df['Global Deaths'], "s-", color=cmap(7))
ax.plot(df.index, df['China Deaths'], "o-", color=cmap(6))

#Drawing grid auxiliary lines
ax.grid()

#Log scale display of y-axis
plt.yscale('log')

#Show legend
plt.legend(['Global Confirmed', 'China Confirmed','Global Deaths', 'China Deaths'])

#Save file
plt.savefig('COVID-19.svg', bbox_inches="tight") #If svg is set to png, it can be saved as an image

The display results are as follows.

Regarding the displayed graph, the horizontal axis is the date and the vertical axis is the number of people.

Finally

• "I want to learn more about AI / machine learning."
• "I want to learn by myself, but I don't know what to learn from."

If you are such a person, you can read this to understand the step-up procedure for learning machine learning efficiently.

• [Self-study] Learning Roadmap for Machine Learning for Python Beginners

Also, here is a reference for specific ways to learn Python.

• [For beginners] Roadmap from the basics of Python to the introduction of machine learning [Free teaching materials released]

Reference information

■ COVID-19 ■ Basic knowledge of matplotlib that I wanted to know early, or the story of an artist who can adjust the appearance