In [3]:
# 基础数据分析库
import numpy as np
import pandas as pd
# 分词库
import jieba
import re
# 绘图库
import matplotlib.pyplot as plt
from pyecharts.charts import *
from pyecharts import options as opts
from pyecharts.charts import Liquid
from pyecharts.globals import ThemeType
import stylecloud
from IPython.display import Image
In [5]:
df = pd.read_csv('messigowhere.csv', names=['评论内容'])
df.head(3)
df.info()
df['评论内容'] = df['评论内容'].astype('str')
In [6]:
result = []
for i in df['评论内容']:
result.append(re.split('[:：,，.。!！~·`\;；……、]',i))
def club_comment(df,result):
clubs = pd.DataFrame(np.zeros(7 * len(df)).reshape(len(df),7),
columns = ['C罗','苏亚雷斯','瓜迪奥拉','内马尔','阿圭罗','巴托梅乌','科曼'])
for i in range(len(result)):
words = result[i]
for word in words:
if '梅西' in word or '苏牙' in word:
clubs.iloc[i + 1]['苏亚雷斯'] = 1
if 'C罗' in word or 'C' in word or 'c' in word or '梅罗' in word:
clubs.iloc[i + 1]['C罗'] = 1
if '苏亚雷斯' in word or '苏牙' in word:
clubs.iloc[i + 1]['苏亚雷斯'] = 1
if '迪奥拉' in word or '瓜帅' in word or '瓜秃' in word:
clubs.iloc[i + 1]['瓜迪奥拉'] = 1
if '内马尔' in word or '马尔' in word:
clubs.iloc[i + 1]['内马尔'] = 1
if '阿圭罗' in word or '阿坤' in word:
clubs.iloc[i + 1]['阿圭罗'] = 1
if '巴托' in word or '大熊' in word:
clubs.iloc[i + 1]['巴托梅乌'] = 1
if '科曼' in word:
clubs.iloc[i + 1]['科曼'] = 1
final_result = pd.concat([df,clubs],axis = 1)
return final_result
In [7]:
df1 = club_comment(df,result)
df1 = df1.drop(labels = 0)
df1.head(5)
Out[7]:
In [8]:
df3 = df1.iloc[:,1:].sum().reset_index().sort_values(0, ascending = True)
df3.columns = ['相关人物','次数']
df3['占比'] = df3['次数'] / df3['次数'].sum()
c = (
Bar(init_opts=opts.InitOpts(theme=ThemeType.WESTEROS))
.add_xaxis(df3['相关人物'].tolist())
.add_yaxis("",df3['次数'].tolist()).reversal_axis() # X轴与y轴调换顺序
.set_global_opts(title_opts=opts.TitleOpts(title="相关人物被提及次数",subtitle="数据来源：直播吧",pos_top="2%",pos_left = 'center'),
xaxis_opts=opts.AxisOpts(axislabel_opts=opts.LabelOpts(font_size=16)), # 更改横坐标字体大小
yaxis_opts=opts.AxisOpts(axislabel_opts=opts.LabelOpts(font_size=16)), # 更改纵坐标字体大小
)
.set_series_opts(label_opts=opts.LabelOpts(font_size=16,position='right'))
)
c.render_notebook()
Out[8]:
In [23]:
messinum = len(df1[df1['评论内容'].str.contains('梅西')])
print(messinum)
messinum_f = round(messinum / len(df1), 2)
m = (
Liquid()
.add("lq", [messinum_f])
.set_global_opts(title_opts=opts.TitleOpts(title="“梅西”一词在评论中出现的频率"))
)
m.render_notebook()
Out[23]:
In [25]:
stopwords = open('baidu_stopwords.txt', encoding='utf-8').read().split('\n')
whole_comment = "".join(df1['评论内容'])
word_list = jieba.cut(whole_comment)
stylecloud.gen_stylecloud(text = ' '.join(word_list), max_words = 1000,
collocations = False,
font_path = 'C:\Windows\Fonts\simkai.ttf',
icon_name = 'fas fa-futbol',
size = 653,
palette = 'matplotlib.Inferno_9',
output_name = './comment.png',
custom_stopwords = stopwords)
Image(filename = './comment.png')
Out[25]:
In [26]:
result = []
for i in df['评论内容']:
result.append(re.split('[:：,，.。!！~·`\;；……、]',i))
def club_comment(df,result):
clubs = pd.DataFrame(np.zeros(8 * len(df)).reshape(len(df),8),
columns = ['曼城','尤文','巴黎','国米','曼联','阿森纳','米兰','切尔西'])
for i in range(len(result)):
words = result[i]
for word in words:
if '曼城' in word:
clubs.iloc[i + 1]['曼城'] = 1
if '尤文' in word:
clubs.iloc[i + 1]['尤文'] = 1
if '巴黎' in word:
clubs.iloc[i + 1]['巴黎'] = 1
if '国米' in word:
clubs.iloc[i + 1]['国米'] = 1
if '曼联' in word:
clubs.iloc[i + 1]['曼联'] = 1
if '阿森纳' in word or '厂' in word:
clubs.iloc[i + 1]['阿森纳'] = 1
if '米兰' in word or '毛' in word:
clubs.iloc[i + 1]['米兰'] = 1
if '切尔西' in word or '车' in word:
clubs.iloc[i + 1]['切尔西'] = 1
final_result = pd.concat([df,clubs],axis = 1)
return final_result
In [27]:
df3 = df1.iloc[:,1:].sum().reset_index().sort_values(0, ascending = True)
df3.columns = ['俱乐部','次数']
df3['占比'] = df3['次数'] / df3['次数'].sum()
c = (
Bar(init_opts=opts.InitOpts(theme=ThemeType.WESTEROS))
.add_xaxis(df3['俱乐部'].tolist())
.add_yaxis("",df3['次数'].tolist()).reversal_axis() # X轴与y轴调换顺序
.set_global_opts(title_opts=opts.TitleOpts(title="俱乐部被提及次数",subtitle="数据来源：直播吧",pos_top="2%",pos_left = 'center'),
xaxis_opts=opts.AxisOpts(axislabel_opts=opts.LabelOpts(font_size=16)), # 更改横坐标字体大小
yaxis_opts=opts.AxisOpts(axislabel_opts=opts.LabelOpts(font_size=16)), # 更改纵坐标字体大小
)
.set_series_opts(label_opts=opts.LabelOpts(font_size=16,position='right'))
)
c.render_notebook()
Out[27]:
In [ ]: