YouTube資料：誰獲得了最多訂閱者？

原文來源：
https://www.kaggle.com/roshan77/youtube-data-who-got-the-most-subscribers
介紹：
Python筆記 使用來自Socialblade的Youtube前5000個頻道的資料。
資料來源：
Socialblade提供的Youtube最受歡迎的5000個頻道資料。
Socialblade列出的YouTube前5000個頻道的總體指標。
https://www.kaggle.com/mdhrumil/top-5000-youtube-channels-data-from-socialblade
資料下載：https://www.kaggle.com/mdhrumil/top-5000-youtube-channels-data-from-socialblade/downloads/top-5000-youtube-channels-data-from-socialblade.zip/2
你對以下方面感興趣:

1、哪個YouTube頻道擁有最多的觀眾、訂閱者、視訊上傳量等等?

2、更多的視訊上傳會帶來更多的視訊瀏覽量和更多的訂閱者嗎?

3、更多的使用者會提供更多的視訊瀏覽量嗎?

4、有沒有一種方法可以根據頻道上傳的視訊數量和觀看視訊的次數來預測訂閱使用者的數量?

我試著用一些視覺工具和分析工具來回答這些問題。

import numpy as np 
import pandas as pd 

import os
print(os.listdir("../input"))  #選擇資料存放路徑 獲取到data.csv檔案

[‘data.csv’]

import matplotlib.pyplot as plt
import seaborn as sns 
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn import metrics
import statsmodels.api as sm
from pandas.core import datetools

df = pd.read_csv('../input/data.csv')

df.head()

Out():
Rank	Grade	Channel name	Video Uploads	Subscribers		Video views
1st		A++		Zee TV			82757			18752951		20869786591
2nd		A++		T-Series		12661			61196302		47548839843
3rd		A++		Cocomelon-Nursery Rhymes	373	19238251		9793305082
4th		A++		SET India		27323			31180559			22675948293
5th		A++		WWE				36756			32852346			26273668433

df.tail()

Out():
		Rank	Grade	Channel name	Video Uploads	Subscribers	Video views
4995	4,996th	B+	Uras Benlioğlu	706	2072942	441202795
4996	4,997th	B+	HI-TECH MUSIC LTD	797	1055091	377331722
4997	4,998th	B+	Mastersaint	110	3265735	311758426
4998	4,999th	B+	Bruce McIntosh	3475	32990	14563764
4999	5,000th	B+	SehatAQUA	254	21172	73312511

df.info()
Out():
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 5000 entries, 0 to 4999
Data columns (total 6 columns):
Rank             5000 non-null object
Grade            5000 non-null object
Channel name     5000 non-null object
Video Uploads    5000 non-null object
Subscribers      5000 non-null object
Video views      5000 non-null int64
dtypes: int64(1), object(5)
memory usage: 234.5+ KB


#df['Subscribers'] = df['Subscribers'].convert_objects(convert_numeric=True)
#df['Video Uploads'] = df['Video Uploads'].convert_objects(convert_numeric=True)

df['Subscribers'] = pd.to_numeric(df['Subscribers'], errors='coerce')
df['Video Uploads'] = pd.to_numeric(df['Video Uploads'], errors='coerce')

探索性資料分析
在這裡，我首先繪製一些條形圖，顯示每種通道分類的前20名。根據排名，前三名都是前20名，他們的上傳視訊數量、訂閱者和視訊瀏覽量都在前三名。根據每個小組各自的情況，最後三名都進入了前20名。

df.head(20).plot.bar(x = 'Channel name', y = 'Subscribers')
plt.title('Number of subscribers of top 20 channels')

df.head(20).plot.bar(x = 'Channel name', y = 'Video views')
plt.title('Number of video views of top 20 channels')

df.head(20).plot.bar(x = 'Channel name', y = 'Video Uploads')
plt.title('Number of video uploads of top 20 channels')

df.sort_values(by = ['Subscribers'], ascending = False).head(20).plot.bar(x = 'Channel name', y = 'Subscribers')
plt.title('Top 20 channels with maximum number of subscribers')

df.sort_values(by = ['Video views'], ascending = False).head(20).plot.bar(x = 'Channel name', y = 'Video views')
plt.title('Top 20 channels with maximum number of video views')

df.sort_values(by = ['Video Uploads'], ascending = False).head(20).plot.bar(x = 'Channel name', y = 'Video Uploads')
plt.title('Top 20 channels with maximum number of video uploads')

在這裡，我感興趣的是列表中的所有通道是如何按照訂閱伺服器、視訊上傳和訂閱伺服器在每個類中從最大值到最小值分佈的。令人感興趣的是，在頂部列表處有巨大的峰值，並且趨向於快速地獲得一個趨於平緩的平臺。

df.sort_values(by = ['Subscribers'], ascending = False).plot(x = 'Channel name', y = 'Subscribers')
plt.xlabel('Ranking by subscribers')
plt.ylabel('Number of subscribers')

df.sort_values(by = ['Video views'], ascending = False).plot(x = 'Channel name', y = 'Video views')
plt.xlabel('Ranking by video views')
plt.ylabel('Number of video views')

df.sort_values(by = ['Video Uploads'], ascending = False).plot(x = 'Channel name', y = 'Video Uploads')
plt.xlabel('Ranking by video uploads')
plt.ylabel('Number of video uploads')

按照頻道等級分析

grade_name = list(set(df['Grade']))
grade_name

output：['A- ', 'A+ ', 'A++ ', '\xa0 ', 'B+ ', 'A ']

df_by_grade = df.set_index(df['Grade'])

count_grade = list()
for grade in grade_name:
    count_grade.append(len(df_by_grade.loc[[grade]]))

df_by_grade.head()

out:
Grade	Rank	Grade	Channel name	Video Uploads	Subscribers	Video views				
A++	1st	A++	Zee TV	82757.0	18752951.0	20869786591
A++	2nd	A++	T-Series	12661.0	61196302.0	47548839843
A++	3rd	A++	Cocomelon - Nursery Rhymes	373.0	19238251.0	9793305082
A++	4th	A++	SET India	27323.0	31180559.0	22675948293
A++	5th	A++	WWE	36756.0	32852346.0	26273668433

print(count_grade)
print(grade_name)

out:
[963, 2956, 10, 1024, 41, 6]
['A ', 'B+ ', 'A++ ', 'A- ', 'A+ ', '\xa0 ']

grade_name[2] = 'missing'

labels = grade_name
sizes = count_grade
explode1 = (0.2, 0.2, 0.5, 0.2, 0.2, 0.2)
color_list = ['green',  'red', 'gold', 'blue', 'lightskyblue', 'brown']
patches, texts = plt.pie(sizes, colors = color_list, explode = explode1, 
                         shadow = False, startangle = 90, radius = 3)
plt.legend(patches, labels, loc = "best")
plt.axis('equal')
plt.title('Classification of channels by grades')
plt.show()

df.describe()
out:

	Video Uploads	Subscribers	Video views
count	4994.000000	4.613000e+03	5.000000e+03
mean	3859.463556	2.620004e+06	1.071449e+09
std	17085.866498	3.926447e+06	2.003844e+09
min	1.000000	3.010000e+02	7.500000e+01
25%	141.000000	6.098940e+05	1.862329e+08
50%	443.000000	1.350477e+06	4.820548e+08
75%	1501.750000	2.950056e+06	1.124368e+09
max	422326.000000	6.599531e+07	4.754884e+10

變數之間的關係

從下圖可以看出，訂閱者的數量與觀看者的數量呈正相關關係。這是預期。但使用者數量與該頻道上傳的視訊數量呈負相關。這可能令人驚訝。吸引更多瀏覽者的視訊頻道和使用者上傳的視訊數量都在減少。