[혼공파] 5주차_혼공분석

 

 

 

 

5. 데이터 시각화하기¶

 

 

5-1. 맷플롯립 기본 요소 알아보기¶

 

 

- Figure 객체¶

 

In [5]:

import pandas as pd

ns_book7 = pd.read_csv('ns_book7.csv', low_memory=False)
ns_book7.head()

 

Out[5]:

	번호	도서명	저자	출판사	발행년도	ISBN	세트 ISBN	부가기호	권	주제분류번호	도서권수	대출건수	등록일자
0	1	인공지능과 흙	김동훈 지음	민음사	2021	9788937444319	NaN	NaN	NaN	NaN	1	0	2021-03-19
1	2	가짜 행복 권하는 사회	김태형 지음	갈매나무	2021	9791190123969	NaN	NaN	NaN	NaN	1	0	2021-03-19
2	3	나도 한 문장 잘 쓰면 바랄 게 없겠네	김선영 지음	블랙피쉬	2021	9788968332982	NaN	NaN	NaN	NaN	1	0	2021-03-19
3	4	예루살렘 해변	이도 게펜 지음, 임재희 옮김	문학세계사	2021	9788970759906	NaN	NaN	NaN	NaN	1	0	2021-03-19
4	5	김성곤의 중국한시기행 : 장강·황하 편	김성곤 지음	김영사	2021	9788934990833	NaN	NaN	NaN	NaN	1	0	2021-03-19

 

In [7]:

import matplotlib.pyplot as plt

plt.scatter(ns_book7['도서권수'], ns_book7['대출건수'], alpha=0.1)
plt.show()

 

 

 

In [8]:

plt.figure(figsize=(9,6))
plt.scatter(ns_book7['도서권수'], ns_book7['대출건수'], alpha=0.1)
plt.show()

 

 

 

In [9]:

print(plt.rcParams['figure.figsize'])

 

 

[6.4, 4.8]

 

In [10]:

print(plt.rcParams['figure.dpi'])

 

 

100.0

 

In [11]:

plt.figure(figsize=(900/100, 600/100))
plt.scatter(ns_book7['도서권수'], ns_book7['대출건수'], alpha=0.1)
plt.show()

 

 

 

In [12]:

%config InlineBackend.print_figure_kwargs = {'bbox_inches': None}
plt.figure(figsize=(900/100, 600/100))
plt.scatter(ns_book7['도서권수'], ns_book7['대출건수'], alpha=0.1)
plt.show()

 

 

 

In [13]:

%config InlineBackend.print_figure_kwargs = {'bbox_inches': 'tight'}

 

In [14]:

plt.figure(dpi=100)
plt.scatter(ns_book7['도서권수'], ns_book7['대출건수'], alpha=0.1)
plt.show()

 

 

 

In [15]:

plt.figure(dpi=200)
plt.scatter(ns_book7['도서권수'], ns_book7['대출건수'], alpha=0.1)
plt.show()

 

 

 

 

- rcParams 객체¶

 

In [18]:

plt.rcParams['figure.dpi'] = 72

 

In [22]:

plt.rcParams['scatter.marker']

 

Out[22]:

'o'

 

In [23]:

plt.rcParams['scatter.marker'] = '*'

 

In [25]:

plt.scatter(ns_book7['도서권수'], ns_book7['대출건수'], alpha=0.1)
plt.show()

 

 

 

In [26]:

plt.scatter(ns_book7['도서권수'], ns_book7['대출건수'], alpha=0.1, marker='+')
plt.show()

 

 

 

 

- 여러 개의 서브플롯 출력하기¶

 

In [30]:

plt.rcParams['figure.dpi'] = 100

 

In [31]:

fig, axs = plt.subplots(2)

axs[0].scatter(ns_book7['도서권수'], ns_book7['대출건수'], alpha=0.1)

axs[1].hist(ns_book7['대출건수'], bins=100)
axs[1].set_yscale('log')

fig.show()

 

 

/var/folders/3w/0y55k0y53pg1dvg3p_cqn81m0000gn/T/ipykernel_29404/3538463139.py:8: UserWarning: FigureCanvasAgg is non-interactive, and thus cannot be shown
  fig.show()

 

 

In [33]:

fig, axs = plt.subplots(2, figsize=(6, 8))

axs[0].scatter(ns_book7['도서권수'], ns_book7['대출건수'], alpha=0.1)
axs[0].set_title('scatter plot')

axs[1].hist(ns_book7['대출건수'], bins=100)
axs[1].set_title('histogram')
axs[1].set_yscale('log')

fig.show()

 

 

/var/folders/3w/0y55k0y53pg1dvg3p_cqn81m0000gn/T/ipykernel_29404/2025912959.py:10: UserWarning: FigureCanvasAgg is non-interactive, and thus cannot be shown
  fig.show()

 

 

In [34]:

fig, axs = plt.subplots(1, 2, figsize=(10,4))

axs[0].scatter(ns_book7['도서권수'], ns_book7['대출건수'], alpha=0.1)
axs[0].set_title('scatter plot')
axs[0].set_xlabel('nnumber of books')
axs[0].set_ylabel('borrow count')

axs[1].hist(ns_book7['대출건수'], bins=100)
axs[1].set_title('histogram')
axs[1].set_yscale('log')
axs[1].set_xlabel('borrow count')
axs[1].set_ylabel('frequency')

fig.show()

 

 

/var/folders/3w/0y55k0y53pg1dvg3p_cqn81m0000gn/T/ipykernel_29404/4022169940.py:14: UserWarning: FigureCanvasAgg is non-interactive, and thus cannot be shown
  fig.show()

 

 

 

5-2. 선 그래프와 막대 그래프 그리기¶

 

 

- 연도별 발행 도서 개수 구하기¶

 

In [37]:

import pandas as pd
ns_book7 = pd.read_csv('ns_book7.csv', low_memory=False)
ns_book7.head()

 

Out[37]:

	번호	도서명	저자	출판사	발행년도	ISBN	세트 ISBN	부가기호	권	주제분류번호	도서권수	대출건수	등록일자
0	1	인공지능과 흙	김동훈 지음	민음사	2021	9788937444319	NaN	NaN	NaN	NaN	1	0	2021-03-19
1	2	가짜 행복 권하는 사회	김태형 지음	갈매나무	2021	9791190123969	NaN	NaN	NaN	NaN	1	0	2021-03-19
2	3	나도 한 문장 잘 쓰면 바랄 게 없겠네	김선영 지음	블랙피쉬	2021	9788968332982	NaN	NaN	NaN	NaN	1	0	2021-03-19
3	4	예루살렘 해변	이도 게펜 지음, 임재희 옮김	문학세계사	2021	9788970759906	NaN	NaN	NaN	NaN	1	0	2021-03-19
4	5	김성곤의 중국한시기행 : 장강·황하 편	김성곤 지음	김영사	2021	9788934990833	NaN	NaN	NaN	NaN	1	0	2021-03-19

 

In [41]:

count_by_year = ns_book7['발행년도'].value_counts()
count_by_year

 

Out[41]:

발행년도
2012    18601
2014    17797
2009    17611
2011    17523
2010    17503
        ...  
2650        1
2108        1
2104        1
2560        1
1947        1
Name: count, Length: 87, dtype: int64

 

In [43]:

count_by_year = count_by_year.sort_index()
count_by_year

 

Out[43]:

발행년도
1947     1
1948     1
1949     1
1952    11
1954     1
        ..
2551     1
2552     2
2559     1
2560     1
2650     1
Name: count, Length: 87, dtype: int64

 

In [45]:

count_by_year = count_by_year[count_by_year.index <= 2030]
count_by_year

 

Out[45]:

발행년도
1947        1
1948        1
1949        1
1952       11
1954        1
        ...  
2020    11834
2021     1255
2025        1
2028        1
2030        1
Name: count, Length: 68, dtype: int64

 

 

- 주제별 도서 개수 구하기¶

 

In [49]:

import numpy as np

def kdc_1st_char(no):
    if no is np.nan:
        return '-1'
    else:
        return no[0]

count_by_subject = ns_book7['주제분류번호'].apply(kdc_1st_char).value_counts()
count_by_subject

 

Out[49]:

주제분류번호
8     108643
3      80767
5      40916
9      26375
6      25070
1      22647
-1     16978
7      15836
4      13688
2      13474
0      12376
Name: count, dtype: int64

 

 

- 선 그래프 그리기¶

 

In [51]:

import matplotlib.pyplot as plt
print(plt.rcParams['figure.dpi'])

 

 

100.0

 

In [52]:

plt.plot(count_by_year.index, count_by_year.values)
plt.title('Books by year')
plt.xlabel('year')
plt.ylabel('number of books')
plt.show()

 

 

 

In [54]:

plt.plot(count_by_year, marker='.', linestyle='--', color='red')
plt.title('Books by year')
plt.xlabel('year')
plt.ylabel('number of books')
plt.show()

 

 

 

In [55]:

plt.plot(count_by_year, '.-.c')

 

Out[55]:

[<matplotlib.lines.Line2D at 0x2848c5c40>]

 

 

In [56]:

plt.plot(count_by_year, '*-g')
plt.title('Books by year')
plt.xlabel('year')
plt.ylabel('number of books')
plt.show()

 

 

 

In [57]:

plt.plot(count_by_year, '*-g')
plt.title('Books by year')
plt.xlabel('year')
plt.ylabel('number of books')
plt.xticks(range(1947, 2030, 10))
for idx, val in count_by_year[::5].items():
    plt.annotate(val, (idx, val))
plt.show()

 

 

 

In [58]:

plt.plot(count_by_year, '*-g')
plt.title('Books by year')
plt.xlabel('year')
plt.ylabel('number of books')
plt.xticks(range(1947, 2030, 10))
for idx, val in count_by_year[::5].items():
    plt.annotate(val, (idx,val), xytext=(idx+1, val+10))
plt.show()

 

 

 

In [60]:

plt.plot(count_by_year, '*-g')
plt.title('Books by year')
plt.xlabel('year')
plt.ylabel('number of books')
plt.xticks(range(1947, 2030, 10))
for idx, val in count_by_year[::5].items():
    plt.annotate(val, (idx,val), xytext=(0, 2), textcoords='offset points', ha='center')
plt.show()

 

 

 

 

- 막대 그래프 그리기¶

 

In [62]:

plt.bar(count_by_subject.index, count_by_subject.values)
plt.title('Books by subject')
plt.xlabel('subject')
plt.ylabel('number of books')
for idx, val in count_by_subject.items():
    plt.annotate(val, (idx,val), xytext=(0, 2), textcoords='offset points')
plt.show()

 

 

 

In [63]:

plt.bar(count_by_subject.index, count_by_subject.values, width=0.7, color='blue')
plt.title('Books by subject')
plt.xlabel('subject')
plt.ylabel('number of books')
for idx, val in count_by_subject.items():
    plt.annotate(val, (idx,val), xytext=(0, 2), textcoords='offset points', fontsize=8, ha='center', color='green')
plt.show()

 

 

 

In [64]:

plt.barh(count_by_subject.index, count_by_subject.values, height=0.7, color='blue')
plt.title('Books by subject')
plt.xlabel('number of books')
plt.ylabel('subject')
for idx, val in count_by_subject.items():
    plt.annotate(val, (val, idx), xytext=(2, 0), textcoords='offset points', fontsize=8, va='center', color='green')
plt.show()

 

 

 

 

- 이미지 출력하고 저장하기¶

 

In [67]:

img = plt.imread('jupiter.png')
img.shape

 

Out[67]:

(1561, 1646, 3)

 

In [68]:

plt.imshow(img)
plt.show()

 

 

 

In [69]:

plt.figure(figsize=(8,6))
plt.imshow(img)
plt.axis('off')
plt.show()

 

 

 

In [70]:

from PIL import Image

pil_img = Image.open('jupiter.png')
plt.figure(figsize=(8,6))
plt.imshow(pil_img)
plt.axis('off')
plt.show()

 

 

 

In [71]:

import numpy as np

arr_img = np.array(pil_img)
arr_img.shape

 

Out[71]:

(1561, 1646, 3)

 

In [76]:

plt.imsave('jupiter.jpg', arr_img)

 

 

- 그래프를 이미지로 저장하기¶

 

In [83]:

plt.rcParams['savefig.dpi']

 

Out[83]:

'figure'

 

In [86]:

plt.barh(count_by_subject.index, count_by_subject.values, height=0.7, color='blue')
plt.title('Books by subject')
plt.xlabel('number of books')
plt.ylabel('subject')
for idx, val in count_by_subject.items():
    plt.annotate(val, (val, idx), xytext=(2, 0), textcoords='offset points', fontsize=8, va='center', color='green')
plt.savefig('books_by_subject.png')
plt.show()

 

 

 

In [89]:

pil_img = Image.open('books_by_subject.png')

plt.figure(figsize=(8,6))
plt.imshow(pil_img)
plt.axis('off')
plt.show()