-
Notifications
You must be signed in to change notification settings - Fork 0
/
report.py
253 lines (208 loc) · 11.6 KB
/
report.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
import pandas as pd
import numpy as np
# from pandas_profiling import ProfileReport
from datetime import datetime
import plotly.express as px
from src.funcs import processed_path, reports_path, relabeling_dict, px_select_deselect
from src.funcs import importDFdtypes
# =============================================================================
# Load and prepare video data
# =============================================================================
# Import csv and assign correct dtypes (both videos and comments)
videos = pd.read_csv(processed_path.joinpath("videos.csv"),
lineterminator="\r",
index_col="videoId")
videos = videos.astype(importDFdtypes("videos.json"))
# Do not use comment_id as index here (replies have same comment_id as the comments they reply to)
comments = pd.read_csv(processed_path.joinpath("comments.csv"),
lineterminator="\r",
index_col=0)
comments = comments.astype(importDFdtypes("comments.json"))
# Only videos published before report_deadline are included in report.
# (to avoid including videos with insufficient time to accumulate comments
report_deadline = "2023-01-01 00:00:00+00:00"
report_deadline_short = "2022"
# =============================================================================
# Video Filter:
# 1) Minimum comments
# 2) only comments before report_deadline
# =============================================================================
min_comments = 50
videos_cutoff = (videos.query("publishedAt < @report_deadline")
.query("available_comments >= @min_comments"))
info_of_used_filter = f'Nur Videos mit min. {min_comments} Kommentaren und bis {report_deadline_short} | {len(videos_cutoff)} von insg. {len(videos)} Videos'
# =============================================================================
# ProfileReport (only for videos, comments too big!)
# =============================================================================
# profile = ProfileReport(videos_cutoff, title="Pandas Profiling Report")
# profile.to_file(reports_path.joinpath("videos_report.html"))
# =============================================================================
# Time series of comment amount
# =============================================================================
frequency = "W" # choose frequency D, W, M, Q
monthly_comments = (comments.query("comment_published < @report_deadline")
.groupby(pd.Grouper(key = "comment_published", freq=frequency))
.size())
timeseries_comments = px.line(monthly_comments,
template = "simple_white")
timeseries_comments.write_html(reports_path.joinpath("timeseries_comments.html"))
# =============================================================================
# Comment feature distributions (splitted by channel)
# NOTE: can also be splitted by category
# =============================================================================
features = ["toplevel_sentiment_mean", "mod_activity", "responsivity", "ratio_RepliesToplevel",
"mean_word_count", "comments_per_author", "removed_comments_perc", "toplevel_neutrality"]
for feature in features:
distributions_allVideos = px.strip(videos_cutoff,
x = feature, y = "videoOwnerChannelTitle",
color = "videoOwnerChannelTitle",
hover_data = ["publishedAt",
"Title",
"n_toplevel_user_comments",
"viewCount",
feature],
template = "simple_white",
#opacity = 0.5,
title = f'ZDF YouTube-Videos | {info_of_used_filter}',
labels = relabeling_dict)
#distributions_allVideos.update_layout(px_select_deselect)
#distributions_allVideos.update_layout(xaxis={'range': [0, 1]})
distributions_allVideos.write_html(reports_path.joinpath(f"Verteilung_{feature}.html"))
# =============================================================================
# SPLOM: Scatter Plot Matrix
# =============================================================================
splom_title = f'Scatterplot Matrix verschiedener YT-Video Eigenschaften | {info_of_used_filter})'
video_features = [#"likes_per_1kViews",
#"mod_activity",
"ratio_RepliesToplevel",
#"responsivity",
#"viewCount",
#"toplevel_sentiment_mean",
"polarity",
"comments_per_author"]
splom = px.scatter_matrix(videos_cutoff,
dimensions = video_features,
#color = "videoOwnerChannelTitle",
hover_data= ["Title"],
template = "simple_white",
opacity = 0.2,
title = splom_title,
labels = relabeling_dict)
splom.update_layout(px_select_deselect)
splom.write_html(reports_path.joinpath("SPLOM.html"))
# Relationship between performance and sentiment-index?
r_squared_matrix_all = (videos_cutoff.corr(numeric_only= True)**2)
r_squared_matrix = (videos_cutoff[video_features].corr(numeric_only=True)**2)
round(r_squared_matrix["polarity"], 2)
# =============================================================================
# Single SCATTER PLOT
# sentiment vs. video kpi
# =============================================================================
kpi = "viewCount"
scatter_plot = px.scatter(videos_cutoff,
y = "toplevel_sentiment_mean", x = kpi,
#facet_col = "videoOwnerChannelTitle",
#facet_col_wrap=5,
#size = "commentCount", # Maybe viewCount here?
#size_max = 55,
hover_data = ["Title", "n_toplevel_user_comments","viewCount", "duration"],
template = "simple_white",
opacity = 0.8,
title = f'ZDF YT-Videos | Beliebtheit vs. Sentiment-Index (Kreisgröße = Views; nur Videos mit mind. {min_comments} Kommentaren und Veröffentlichung vor 1.8.22, {info_of_used_filter})',
labels = relabeling_dict)
scatter_plot.update_layout(px_select_deselect)
scatter_plot.write_html(reports_path.joinpath("scatter_plot.html"))
sentiment_vs_popularity = videos_cutoff[["videoOwnerChannelTitle", kpi, "toplevel_sentiment_mean"]]
for channel in sentiment_vs_popularity["videoOwnerChannelTitle"].unique():
r_matrix = sentiment_vs_popularity[sentiment_vs_popularity["videoOwnerChannelTitle"] == channel].corr(numeric_only= True)
r_squared_matrix = round(r_matrix ** 2, 3)
print(channel, r_squared_matrix.iloc[0,1])
# =============================================================================
# Channels quarterly resolution
# =============================================================================
# Generate new path for quarterly reports
quarter_path = reports_path.joinpath("Quartalszahlen")
quarter_path.mkdir(exist_ok = True)
# Augment quarter (as float .1 = Q1, .2 = Q2, etc.)
videos_cutoff["quarter"] = (
videos_cutoff["publishedAt"].dt.year +
(videos_cutoff["publishedAt"].dt.quarter / 10)
)
comments["quarter"] = (
comments["comment_published"].dt.year +
(comments["publishedAt"].dt.quarter / 10)
.apply(lambda x: round(x, 1))
)
# Loop through quarters and aggregate metrics
quarters = sorted(list(videos_cutoff["quarter"].unique()))
channels_quarter = pd.DataFrame()
for quarter in quarters:
video_derived_metrics = (
videos_cutoff
.query("quarter == @quarter")
.groupby("videoOwnerChannelTitle")
.agg({
"quarter" : "median",
"video_url" : "size", # column only used to sum videos (relabeled below)
"viewCount": "sum",
"likeCount" : "sum",
"commentCount" : "sum",
"available_comments": "sum",
"removed_comments" : "sum",
"removed_comments_perc" : "mean",
"comments_per_author" : "mean",
"ratio_RepliesToplevel" : "mean",
"toplevel_neutrality": "mean",
"responsivity" : "mean",
"toplevel_sentiment_mean": "mean",
})
).reset_index()
comment_derived_metrics = (
comments
.query("quarter == @quarter")
.groupby("videoOwnerChannelTitle")
.agg({
"videoId" : "size", # column only used to sum comments (relabeled below)
"comment_word_count" : "median",
"owner_comment" : "sum",
})
).reset_index().drop("videoOwnerChannelTitle", axis = 1)
derived_metrics = pd.concat([video_derived_metrics, comment_derived_metrics], axis = 1)
channels_quarter = pd.concat([channels_quarter, derived_metrics], axis = 0)
print(f"estimated metrics for {quarter}")
# Clean up dataframe
channels_quarter["videoCount"] = channels_quarter["video_url"]
channels_quarter["commentCount"] = channels_quarter["videoId"]
channels_quarter["mod_activity"] = channels_quarter["owner_comment"] / channels_quarter["available_comments"] * 1000
channels_quarter = channels_quarter.drop(["video_url", "videoId"], axis = 1).reset_index(drop = True)
channels_quarter["quarter_cat"] = pd.Categorical(channels_quarter["quarter"], categories=quarters, ordered=True)
channels_quarter["quarter_cat"] = channels_quarter["quarter_cat"].cat.rename_categories(lambda x: str(x).replace(".", " Q"))
channels_quarter = channels_quarter.dropna()
# Export table
channels_quarter = channels_quarter.sort_values(["videoOwnerChannelTitle", "quarter"]).reset_index(drop=True)
(channels_quarter.rename(columns = relabeling_dict)
.to_csv(processed_path.joinpath("Quartalszahlen.csv"),
lineterminator="\r",
index = False))
channels_quarter["publishedAt"] = channels_quarter["publishedAt"].apply(lambda x: x.tz_localize(None))
(channels_quarter.rename(columns = relabeling_dict)
.to_excel(quarter_path.joinpath("Quartalszahlen.xlsx"),
sheet_name="Quartalszahlen"))
# Quarterly plots (for each feature a single plot)
# min_quarter = 2019.1
# channels_quarter_plot = channels_quarter.query("quarter >= @min_quarter")
features = ["toplevel_sentiment_mean", "responsivity", "mod_activity", "removed_comments_perc"]
for feature in features:
quaterly_metrics = px.line(channels_quarter,
x = "quarter_cat", y = feature,
color = "videoOwnerChannelTitle",
hover_data = ["videoCount"],
template = "simple_white",
#opacity = 0.8,
markers = True,
title = f'ZDF YT-Videos | {relabeling_dict.get(feature)} (Quartalsmittelwerte)',
labels = relabeling_dict)
quaterly_metrics.update_layout(px_select_deselect)
quaterly_metrics.update_xaxes(title = None)
file_name = f"Quartalsverlauf_{relabeling_dict.get(feature).replace(' ','_')}.html"
quaterly_metrics.write_html(quarter_path.joinpath(file_name))