[Text]S8_07_Text_visualization

Echodot2_Reviews.csv

2.34MB

 

 

 

데이터 출처:

https://www.kaggle.com/datasets/PromptCloudHQ/amazon-echo-dot-2-reviews-dataset

 

In [150]:

import pandas as pd
import string
import nltk # natural language processing (자연어 처리)
from nltk.corpus import stopwords
import gensim # 자연어 처리 중 토큰화
from gensim.utils import simple_preprocess

import matplotlib.pyplot as plt
import seaborn as sns

 

In [135]:

echo_df =  pd.read_csv('Echodot2_Reviews.csv', encoding='utf-8')
echo_df.head(2)

 

Out[135]:

	Rating	Review Date	Configuration Text	Review Text	Review Color	Title	User Verified	Review Useful Count	Declaration Text	Pageurl
0	3	10/3/2017	Echo Dot	Not great speakers	Black	Three Stars	Verified Purchase	NaN	NaN	https://www.amazon.com/All-New-Amazon-Echo-Dot...
1	4	9/26/2017	Echo Dot	Great little gagit	White	Four Stars	Verified Purchase	NaN	NaN	https://www.amazon.com/All-New-Amazon-Echo-Dot...

 

In [136]:

echo_df.dropna(subset=['Review Text'], how='any', axis = 0, inplace=True)

 

In [137]:

# A function to remove punctuations
def remove_punc(text):
  Test_removed_punc = [ char for char in text if char not in string.punctuation ]
  Test_punc_join = ''.join(Test_removed_punc)
  return Test_punc_join

 

In [138]:

# download stopwords (불용어) - 감성분석
# 그 자체로 의미가 없는 것들
nltk.download('stopwords')

 

 

[nltk_data] Downloading package stopwords to /root/nltk_data...
[nltk_data]   Package stopwords is already up-to-date!

Out[138]:

True

 

In [139]:

# Add more stopwords : extend
stop_words = stopwords.words('english')
stop_words.extend(['amazon','Amazon','echo','echo','echo','device','Dot','dot'])

 

In [140]:

# Remove stopwords and remove short words(less than 2 char)
def preprocess(text):
  result = []
  for token in gensim.utils.simple_preprocess(text):
    if token not in stop_words and len(token) >= 1:
      result.append(token)
  return result

 

In [141]:

echo_df['reviews_unpunc']= echo_df['Review Text'].apply(remove_punc)
echo_df['reviews_unpunc_unstopwords'] = echo_df['reviews_unpunc'].apply(preprocess)

 

In [142]:

# join the words into a string
echo_df['Adj_text'] = echo_df['reviews_unpunc_unstopwords'].apply(lambda x: ' '.join(x))

 

 

tokenization

 

In [143]:

# tokenization
from sklearn.feature_extraction.text import CountVectorizer

vectorizer = CountVectorizer() # instance
echo_CVec = vectorizer.fit_transform(echo_df['Adj_text'])

 

 

Visualization

 

In [144]:

# the number of words
echo_df['length'] = echo_df['reviews_unpunc_unstopwords'].apply(len)

 

In [145]:

# plot the histogram of the length column
echo_df['length'].hist(bins=100, figsize = (20,10) )

 

Out[145]:

<matplotlib.axes._subplots.AxesSubplot at 0x7efe5b6fb160>

 

 

In [146]:

# positive/negative
def feedback_class(row):
  if '4' in str(row) or '5' in str(row):
    return 1
  else:
    return 0

 

In [147]:

echo_df['feedback'] = echo_df['Rating'].apply(feedback_class)
echo_df.head() 

 

Out[147]:

	Rating	Review Date	Configuration Text	Review Text	Review Color	Title	User Verified	Review Useful Count	Declaration Text	Pageurl	reviews_unpunc	reviews_unpunc_unstopwords	Adj_text	length	feedback
0	3	10/3/2017	Echo Dot	Not great speakers	Black	Three Stars	Verified Purchase	NaN	NaN	https://www.amazon.com/All-New-Amazon-Echo-Dot...	Not great speakers	[great, speakers]	great speakers	2	0
1	4	9/26/2017	Echo Dot	Great little gagit	White	Four Stars	Verified Purchase	NaN	NaN	https://www.amazon.com/All-New-Amazon-Echo-Dot...	Great little gagit	[great, little, gagit]	great little gagit	3	1
2	5	9/8/2017	Echo Dot	Awesome 👏🏽	White	Awesome!	Verified Purchase	NaN	NaN	https://www.amazon.com/All-New-Amazon-Echo-Dot...	Awesome 👏🏽	[awesome]	awesome	1	1
3	5	10/19/2017	Echo Dot	Love my Echo	Black	Five Stars	Verified Purchase	NaN	NaN	https://www.amazon.com/All-New-Amazon-Echo-Dot...	Love my Echo	[love]	love	1	1
4	5	9/17/2017	Echo Dot	Great device	Black	Five Stars	Verified Purchase	NaN	NaN	https://www.amazon.com/All-New-Amazon-Echo-Dot...	Great device	[great]	great	1	1

 

In [148]:

sns.countplot(x='feedback', data = echo_df)

 

Out[148]:

<matplotlib.axes._subplots.AxesSubplot at 0x7efe570a65b0>

 

 

In [149]:

plt.figure(figsize=(25,10))
sns.barplot(x='Configuration Text', y= 'Rating', data = echo_df, palette='deep')
plt.xticks(rotation='45')

 

Out[149]:

(array([0, 1]), <a list of 2 Text major ticklabel objects>)

 

 

In [151]:

sns.countplot(x='Rating', data= echo_df)

 

Out[151]:

<matplotlib.axes._subplots.AxesSubplot at 0x7efe59e8d4c0>