-
Notifications
You must be signed in to change notification settings - Fork 0
/
reader.py
296 lines (256 loc) · 11 KB
/
reader.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
import pandas as pd
from nltk.stem import WordNetLemmatizer
from nltk.corpus import stopwords
import re
from wordcloud import WordCloud
import matplotlib.pyplot as plt
plt.style.use('ggplot')
TEAL = '#95D0F7'
TAN = '#EDDECD'
SILK = '#F2ECE5'
BLUE = '#6E9BB9'
GREY = '#858585'
BLACK = '#000000'
WHITE = '#FFFFFF'
PATH = './visuals/'
class Reader:
# Takes a pandas dataframe reviews to analyze
def __init__(self, reviews):
self._reviews = reviews
temp = pd.read_excel('./content_map.xlsx')
self._map = {key: temp[key].tolist() for key in temp.columns}
# Remove NaN values
for key in self._map.keys():
self._map[key] = [x for x in self._map[key] if isinstance(x, str)]
self._categories = {key: ({-1},{-1}) for key in self._map.keys()}
self._categorize()
def get_categories(self):
return self._categories
def get_categories_pos(self):
pos = {key: self._categories[key][1] for key in self._categories.keys()}
return pos
def get_categories_neg(self):
neg = {key: self._categories[key][0] for key in self._categories.keys()}
return neg
def analyze(self):
#self._count_reviews()
maxes = self._percent_reviews()
self._create_wordcloud(pos=True)
self._create_wordcloud(pos=False)
return maxes
def get_helpful_reviews(self):
'''
Returns a tuple of dictionaries that contain 0=positive,
1=negative reviews for each category
'''
g = self._helpful_reviews(self.get_categories_pos())
n = self._helpful_reviews(self.get_categories_neg())
all_helpful = (g, n)
return all_helpful
def _count_reviews(self):
pos = self.get_categories_pos()
neg = self.get_categories_neg()
cat = pos.keys()
pos_counts = [len(pos[l]) for l in pos]
neg_counts = [len(neg[l]) for l in neg]
fig, ax = plt.subplots(1,1)
plt.bar(cat, pos_counts, label='positive', color=BLUE, bottom=neg_counts)
plt.bar(cat, neg_counts, label='negative', color=PURPLE,)
ax.set_facecolor(SILK)
ax.legend()
plt.savefig(PATH+'review_counts.png')
def _percent_reviews(self):
pos = self.get_categories_pos()
neg = self.get_categories_neg()
cat = pos.keys()
# total review for each category
totals = [(len(pos[c]) + len(neg[c])) for c in cat]
pos_rev = [len(pos[l]) for l in pos]
percentages = {}
i = 0
pos_max = ('default', -1)
neg_max = ('default', -1)
for c in cat:
positive_review_percent = int((pos_rev[i] / totals[i]) * 100)
percentages[c] = [positive_review_percent]
percentages[c].append(100 - positive_review_percent)
i += 1
if positive_review_percent > pos_max[1]:
pos_max = (c, positive_review_percent)
elif (100 - positive_review_percent) > neg_max[1]:
neg_max = (c, 100 - positive_review_percent)
# graphing
# make font larger and change style
font = {'family' : 'Tahoma',
'weight' : 'normal',
'size' : 25}
plt.rc('font', **font)
labels = 'Positive', 'Negative'
fig, [[ax1, ax2, ax3], [ax4, ax5, ax6]] = plt.subplots(2, figsize=(20, 10), ncols=3)
title_color = BLACK
title_size = 25
sizes = percentages['food']
ax1.pie(sizes, labels=labels, shadow=False, autopct='%1.0f%%', startangle=90, labeldistance=None, colors=[TEAL,BLUE])
ax1.axis('equal') # Equal aspect ratio ensures that pie is drawn as a circle.
ax1.set_title('Food Reviews', fontsize=title_size, color=title_color)
ax2.axis('off')
sizes = percentages['brunch']
ax3.pie(sizes, labels=labels, shadow=False, autopct='%1.0f%%', startangle=90, labeldistance=None, colors=[TEAL,BLUE])
ax3.axis('equal') # Equal aspect ratio ensures that pie is drawn as a circle.
ax3.set_title('Brunch Reviews', fontsize=title_size, color=title_color)
sizes = percentages['dinner']
ax4.pie(sizes, labels=labels, shadow=False, autopct='%1.0f%%', startangle=90, labeldistance=None, colors=[TEAL,BLUE])
ax4.axis('equal') # Equal aspect ratio ensures that pie is drawn as a circle.
ax4.set_title('Dinner Reviews', fontsize=title_size, color=title_color)
sizes = percentages['ambiance']
ax5.pie(sizes, labels=labels, shadow=False, autopct='%1.0f%%', startangle=90, labeldistance=None, colors=[TEAL,BLUE])
ax5.axis('equal') # Equal aspect ratio ensures that pie is drawn as a circle.
ax5.set_title('Ambiance Reviews', fontsize=title_size, color=title_color)
sizes = percentages['service']
ax6.pie(sizes, labels=labels, shadow=False, autopct='%1.0f%%', startangle=90, labeldistance=None, colors=[TEAL,BLUE])
ax6.axis('equal') # Equal aspect ratio ensures that pie is drawn as a circle.
ax6.set_title('Service Reviews', fontsize=title_size, color=title_color)
ax1.legend(loc='upper right', bbox_to_anchor=(2.05,1))
fig.patch.set_facecolor(WHITE)
plt.savefig(PATH+'all_piecharts.png', bbox_inches='tight', pad_inches=0)
return (neg_max, pos_max)
def _plot_piechart(self, category, percentages, ax):
#labels = 'Positive', 'Negative'
sizes = percentages[category]
fig1, ax1 = plt.subplots()
ax1.pie(sizes, shadow=True, autopct='%1.1f%%', startangle=90)
ax1.axis('equal') # Equal aspect ratio ensures that pie is drawn as a circle.
plt.title(category + ' review percentages')
plt.savefig(category + 'piechart.png')
# Takes a dictionary of lists reviews to get review content
# Takes boolean pos to determine if wordcloud is of positive or negative reviews
# Takes a string cat that determins which category is displayed. Defaults to all.
def _create_wordcloud(self, pos, cat='all'):
data = {}
stopwords = {'good',
'great',
'nice',
'bad',
'horrible',
'terrible',
'amazing',
'however',
'not',
'no',
'well',
'restaurant',
'definitely',
'go',
'really',
'week',
'would',
'overall',
'try',
'give',
'could',
'u',
'food',
'service',
'brunch'}
label=''
if pos:
data = self.get_categories_pos()
label = 'pos'
else:
data = self.get_categories_neg()
label = 'neg'
# Get all relevant rows
all_rows = set()
if cat == 'all':
for val in data.values():
all_rows = all_rows.union(val)
else:
try:
all_rows = all_rows.union(data[cat])
except KeyError:
print('Invalid input parameter for \'cat\'')
return None
words = ''
for row in all_rows:
words += self._clean(self._reviews.loc[row, 'review'])
wordcloud = WordCloud(stopwords=stopwords, width=1600, height=800, background_color=WHITE, max_words=100).generate(words)
fig1, ax = plt.subplots(1,1)
plt.figure(figsize=(2,1), dpi=400)
plt.imshow(wordcloud, aspect='equal')
plt.axis('off')
plt.tight_layout(pad=0)
plt.savefig(PATH+'wordcloud_'+ cat + '_' + label + '.png', facecolor=None)
def _helpful_reviews(self, reviews):
'''
creates a dictionary of reviews that has the categories
as keys and reviews as the values
'''
review_dict= {}
for c in reviews:
review_dict[c] = []
all_reviews = reviews[c]
for r in all_reviews:
if (len(self._reviews.loc[r, 'review']) > 200):
review_dict[c].append(self._reviews.loc[r, 'review'])
return review_dict
# reviews is a pandas dataframe
def _categorize(self):
p_reviews = self._reviews.copy()
# Clean reviews for easier processing
for i in range(len(p_reviews)):
p_reviews.loc[i, 'review'] = self._clean(p_reviews.loc[i, 'review'])
# Categorize reviews
# for each category of the map:
for c in self._map:
# for each word in the map
for w in self._map[c]:
# for each review in the reviews
for i in range(len(p_reviews)):
# Check if any matches from the reviews
if w in p_reviews.loc[i, 'review']:
# If greater or equal to 3, place in [1] (positive review)
rating = -1
if (c == 'brunch') | (c == 'dinner'):
rating = p_reviews.loc[i, 'food']
else:
rating = p_reviews.loc[i, c]
# If there is a match, check if the rating is >= 3 stars
# If less than 3, place in [0] (negative review)
index = 0
if rating > 3:
index = 1
self._categories[c][index].add(i)
# Also add to brunch/dinner if category is food
if (c == 'brunch') | (c == 'dinner'):
self._categories['food'][index].add(i)
self._categories[c][0].remove(-1)
self._categories[c][1].remove(-1)
'''
Using sentiment analysis
processed_reviews = reviews.copy()
for review in reviews['Review']:
processed_reviews.append(self._clean(review))
p_reviews = pd.DataFrame(processed_reviews)
p_reviews.columns = ['Review']
for category in self._map:
p_review_id = 0
for p_review in p_reviews['Review']:
line = p_review.split(' ')
for word in line:
if self._map[category].str.contains(word).sum() > 0:
self._categories[category].add(p_review_id)
p_review_id += 1
for s in self._categories:
self._categories[s].remove(-1)
'''
# Removes stop words, and lemmatizes given review
def _clean(self, review):
wl = WordNetLemmatizer()
all_stopwords = stopwords.words('english')
all_stopwords.remove('not')
review = re.sub('[^a-zA-Z]', ' ', review)
review = review.lower()
review = review.split()
review = [wl.lemmatize(word) for word in review if not word in all_stopwords]
review = ' '.join(review)
return review