WordCloud（ワードクラウド）

from google.colab import drive
drive.mount('/content/drive')

!apt-get -y install fonts-ipafont-gothic
!pip install Janome

from janome.tokenizer import Tokenizer
from wordcloud import WordCloud
import matplotlib.pyplot as plt 
 
text_file = open("/content/drive/My Drive/Colab Notebooks/〇〇/◇△.txt")
full_text = text_file.read()
full_text= full_text.replace("\n","")
 
#print(full_text)  
t = Tokenizer()
tokens = t.tokenize(full_text)
 
 
word_list=[]
for token in tokens:
    word = token.surface
    partOfSpeech = token.part_of_speech.split(',')[0]
    partOfSpeech2 = token.part_of_speech.split(',')[1]
     
    if partOfSpeech == "名詞":
        if (partOfSpeech2 != "非自立") and (partOfSpeech2 != "代名詞") and (partOfSpeech2 != "数"):
            word_list.append(word)
 
words_wakati=" ".join(word_list)
print(words_wakati)  
 
stop_words = ['さん','そう']  
fpath = '/usr/share/fonts/truetype/fonts-japanese-gothic.ttf'  # 日本語フォント指定
 
wordcloud = WordCloud(
    font_path=fpath,
    width=900, height=600,   # default width=400, height=200
    background_color="white",   # default=”black”
    stopwords=set(stop_words),
    max_words=500,   # default=200
    min_font_size=4,   #default=4
    collocations = False,   #default = True
    #colormap='flag'
    ).generate(words_wakati)
 
plt.figure(figsize=(15,12))
plt.imshow(wordcloud)
plt.axis("off")
plt.savefig("word_cloud.png")
plt.show()