Pandas (4) Dataframe 연산

 

전체 DataFrame에 대한 연산

 

In [31]:

pd.options.display.max_rows = 8
movie = pd.read_csv(r'C:\Users\user\jupyterpractice\EDA\Pandas-Cookbook-master\data\movie.csv')
movie.shape

 

Out[31]:

(4916, 28)

 

 

In [32]:

movie.size

 

Out[32]:

137648

 

 

In [33]:

# numpy에서 ndim은 차원의 수를 반환함
movie.ndim

 

Out[33]:

2

 

 

In [34]:

# len은 row의 수를 반환함
len(movie)

 

Out[34]:

4916

 

 

In [35]:

# count method로 각 열의 누락값을 제외한 실제 값의 개수를 알아낸다.
movie.count()

 

Out[35]:

color                     4897
director_name             4814
num_critic_for_reviews    4867
duration                  4901
                          ... 
actor_2_facebook_likes    4903
imdb_score                4916
aspect_ratio              4590
movie_facebook_likes      4916
Length: 28, dtype: int64

 

 

In [36]:

movie.min()

 

Out[36]:

num_critic_for_reviews        1
duration                      7
director_facebook_likes       0
actor_3_facebook_likes        0
                           ... 
actor_2_facebook_likes        0
imdb_score                  1.6
aspect_ratio               1.18
movie_facebook_likes          0
Length: 19, dtype: object

 

 

In [37]:

# 위의 모든 descriptive statistics(기술적인 통계)를 반환하는 describe() method
# 결과는 descriptive statistics를 index로 가지는 DataFrame
movie.describe()

 

Out[37]:

	num_critic_for_reviews	duration	director_facebook_likes	actor_3_facebook_likes	actor_1_facebook_likes	gross	num_voted_users	cast_total_facebook_likes	facenumber_in_poster	num_user_for_reviews	budget	title_year	actor_2_facebook_likes	imdb_score	aspect_ratio	movie_facebook_likes
count	4867.000000	4901.000000	4814.000000	4893.000000	4909.000000	4.054000e+03	4.916000e+03	4916.000000	4903.000000	4895.000000	4.432000e+03	4810.000000	4903.000000	4916.000000	4590.000000	4916.000000
mean	137.988905	107.090798	691.014541	631.276313	6494.488491	4.764451e+07	8.264492e+04	9579.815907	1.377320	267.668846	3.654749e+07	2002.447609	1621.923516	6.437429	2.222349	7348.294142
std	120.239379	25.286015	2832.954125	1625.874802	15106.986884	6.737255e+07	1.383222e+05	18164.316990	2.023826	372.934839	1.002427e+08	12.453977	4011.299523	1.127802	1.402940	19206.016458
min	1.000000	7.000000	0.000000	0.000000	0.000000	1.620000e+02	5.000000e+00	0.000000	0.000000	1.000000	2.180000e+02	1916.000000	0.000000	1.600000	1.180000	0.000000
25%	49.000000	93.000000	7.000000	132.000000	607.000000	5.019656e+06	8.361750e+03	1394.750000	0.000000	64.000000	6.000000e+06	1999.000000	277.000000	5.800000	1.850000	0.000000
50%	108.000000	103.000000	48.000000	366.000000	982.000000	2.504396e+07	3.313250e+04	3049.000000	1.000000	153.000000	1.985000e+07	2005.000000	593.000000	6.600000	2.350000	159.000000
75%	191.000000	118.000000	189.750000	633.000000	11000.000000	6.110841e+07	9.377275e+04	13616.750000	2.000000	320.500000	4.300000e+07	2011.000000	912.000000	7.200000	2.350000	2000.000000
max	813.000000	511.000000	23000.000000	23000.000000	640000.000000	7.605058e+08	1.689764e+06	656730.000000	43.000000	5060.000000	4.200000e+09	2016.000000	137000.000000	9.500000	16.000000	349000.000000

 

 

In [38]:

pd.options.display.max_rows = 10

 

 

In [39]:

# percentiles 매개변수를 통해 정확한 분위수를 지정할 수 있음
movie.describe(percentiles=[.01, .3, .99])

 

Out[39]:

	num_critic_for_reviews	duration	director_facebook_likes	actor_3_facebook_likes	actor_1_facebook_likes	gross	num_voted_users	cast_total_facebook_likes	facenumber_in_poster	num_user_for_reviews	budget	title_year	actor_2_facebook_likes	imdb_score	aspect_ratio	movie_facebook_likes
count	4867.000000	4901.000000	4814.000000	4893.000000	4909.000000	4.054000e+03	4.916000e+03	4916.000000	4903.000000	4895.000000	4.432000e+03	4810.000000	4903.000000	4916.000000	4590.000000	4916.000000
mean	137.988905	107.090798	691.014541	631.276313	6494.488491	4.764451e+07	8.264492e+04	9579.815907	1.377320	267.668846	3.654749e+07	2002.447609	1621.923516	6.437429	2.222349	7348.294142
std	120.239379	25.286015	2832.954125	1625.874802	15106.986884	6.737255e+07	1.383222e+05	18164.316990	2.023826	372.934839	1.002427e+08	12.453977	4011.299523	1.127802	1.402940	19206.016458
min	1.000000	7.000000	0.000000	0.000000	0.000000	1.620000e+02	5.000000e+00	0.000000	0.000000	1.000000	2.180000e+02	1916.000000	0.000000	1.600000	1.180000	0.000000
1%	2.000000	43.000000	0.000000	0.000000	6.080000	8.474800e+03	5.300000e+01	6.000000	0.000000	1.940000	6.000000e+04	1951.000000	0.000000	3.100000	1.330000	0.000000
30%	60.000000	95.000000	11.000000	176.000000	694.000000	7.914069e+06	1.186450e+04	1684.500000	0.000000	80.000000	8.000000e+06	2000.000000	345.000000	6.000000	1.850000	0.000000
50%	108.000000	103.000000	48.000000	366.000000	982.000000	2.504396e+07	3.313250e+04	3049.000000	1.000000	153.000000	1.985000e+07	2005.000000	593.000000	6.600000	2.350000	159.000000
99%	546.680000	189.000000	16000.000000	11000.000000	44920.000000	3.264128e+08	6.815846e+05	62413.900000	8.000000	1999.240000	2.000000e+08	2016.000000	17000.000000	8.500000	4.000000	93850.000000
max	813.000000	511.000000	23000.000000	23000.000000	640000.000000	7.605058e+08	1.689764e+06	656730.000000	43.000000	5060.000000	4.200000e+09	2016.000000	137000.000000	9.500000	16.000000	349000.000000

 

 

In [40]:

pd.options.display.max_rows = 8

 

 

In [42]:

# 결측치(누락값)의 개수 세기 : isnull().sum() method chaining
movie.isnull().sum()

 

Out[42]:

color                      19
director_name             102
num_critic_for_reviews     49
duration                   15
                         ... 
actor_2_facebook_likes     13
imdb_score                  0
aspect_ratio              326
movie_facebook_likes        0
Length: 28, dtype: int64

 

 

skipna 매개변수 : 결측치(누락값)을 무시하지 않는 방법¶

• pandas는 default로 수치열의 누락값을 무시하고 통계값을 냄 -> skipna=True가 default인 것.
• skipna = False로 해줌으로써, 하나라도 누락값(결측치)가 있으면 NaN을 반환하도록 할 수 있음.

 

In [29]:

movie.min(skipna=False)

 

Out[29]:

num_critic_for_reviews     NaN
duration                   NaN
director_facebook_likes    NaN
actor_3_facebook_likes     NaN
                          ... 
actor_2_facebook_likes     NaN
imdb_score                 1.6
aspect_ratio               NaN
movie_facebook_likes       0.0
Length: 16, dtype: float64