데이스쿨 스터디 - 전력 사용량 예측 [2] (전처리)

➢ sns.pointplot : 중앙값으로 표현하는 plot

➢ 기타. 

1. building_type 이라는 컬럼명에서 type 별로 분류하는 코드. 

other_buildings = train_merge[train_merge['building_type'] == 'Other_Buildings']
Public = train_merge[train_merge['building_type'] == 'Public']
University = train_merge[train_merge['building_type'] == 'University']
Data_Center = train_merge[train_merge['building_type'] == 'Data_Center']
Department_Store_and_Outlet = train_merge[train_merge['building_type'] == 'Department_Store_and_Outlet']
Hospital = train_merge[train_merge['building_type'] == 'Hospital']
Commercial = train_merge[train_merge['building_type'] == 'Commercial']
Apartment = train_merge[train_merge['building_type'] == 'Apartment']
Research_Institute = train_merge[train_merge['building_type'] == 'Research_Institute']
Knowledge_Industry_Center = train_merge[train_merge['building_type'] == 'Knowledge_Industry_Center']
Discount_Mart = train_merge[train_merge['building_type'] == 'Discount_Mart']
Hotel_and_Resort = train_merge[train_merge['building_type'] == 'Hotel_and_Resort']

2. 두개의 excel shit를  하나의 컬럼에 맞게 병합하는 방법.  →  pd.merge 사용  

 

# 설명 : building_num 을 기준으로 병합을 수행함, 
#       how='left'라는 인자로 인해 왼쪽에 있는 train_data기준으로 병합함.

train_merge = pd.merge(train_data, building_info, on='building_num', how='left')
print(train_merge.isnull().sum())

->
building_num             0
data_time                0
temperature              0
windspeed               19
humidity                 9
consumption              0
hour                     0
dow                      0
month                    0
week                     0
day                      0
sin_time                 0
cos_time                 0
holiday                  0
building_type            0
total_area               0
cooling_area             0
solar_power_capacity     0
ess_capacity             0
pcs_capacity             0
dtype: int64

# 데이터 모습

print(train_data.isnull().sum()) 
print()
print(building_info.isnull().sum())

출력모습 ->
building_num     0
data_time        0
temperature      0
windspeed       19
humidity         9
consumption      0
hour             0
dow              0
month            0
week             0
day              0
sin_time         0
cos_time         0
holiday          0
dtype: int64

building_num            0
building_type           0
total_area              0
cooling_area            0
solar_power_capacity    0
ess_capacity            0
pcs_capacity            0
dtype: int64