산업부문 요인분해분석 예시코드

참고사항

  • 광업, 제조업을 19개 업종으로 나누어 에너지, 온실가스 요인분해분석을 수행
  • 한국에너지공단 에너지, 온실가스 통계자료와 한국은행 국민계정(실질산출액, 실질부가가치액)을 활용
  • 해당코드는 python 3.7에서 작성되었습니다.
In [1]:
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline
In [2]:
# 데이터 불러오기
df = pd.read_excel('total_input.xlsx')
df.head()
Out[2]:
실적년도 감축업종코드 감축업종코드명 실질산출액(백만원) 실질부가가치액(백만원) 석탄류(천toe) 석유류(천toe) 가스류(천toe) 기타연료(천toe) 열(천toe) 전력(천toe) 합계(천toe) 석탄류(천tCO2) 석유류(천tCO2) 가스류(천tCO2) 기타연료(천tCO2) 열(천tCO2) 전력(천tCO2) 합계(천tCO2)
0 2011 1 광업 3.796199e+06 2.365625e+06 35.379990 21.930911 0.078128 0.000000 0.107800 96.804241 154.301071 139.436559 70.335699 0.16825 0.000000 0.395267 509.81108 720.146855
1 2011 2 음식료품 9.892679e+07 1.721800e+07 31.338552 356.569321 707.434415 9.210000 157.935973 833.840325 2096.328586 117.007696 987.798086 1510.08146 33.770000 579.098567 4391.34344 7619.099249
2 2011 3 섬유 8.421345e+07 1.786797e+07 79.273859 195.799004 343.080913 11.827000 348.709005 1217.604782 2196.294563 295.517393 576.126473 725.45317 43.365667 1278.599685 6412.40925 9331.471639
3 2011 4 목재 7.154951e+06 2.010102e+06 244.525355 21.946030 160.746631 0.000000 11.580103 119.796444 558.594563 896.607615 65.590916 340.52884 0.000000 42.460377 630.89797 1976.085718
4 2011 5 제지 3.334314e+07 1.062102e+07 52.852203 473.885984 381.283795 287.297308 307.554734 896.013941 2398.887966 197.000381 1590.116088 806.18153 1053.423463 1127.700692 4718.77984 9493.201994
In [3]:
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 171 entries, 0 to 170
Data columns (total 19 columns):
실적년도            171 non-null int64
감축업종코드          171 non-null int64
감축업종코드명         171 non-null object
실질산출액(백만원)      171 non-null float64
실질부가가치액(백만원)    171 non-null float64
석탄류(천toe)       171 non-null float64
석유류(천toe)       171 non-null float64
가스류(천toe)       171 non-null float64
기타연료(천toe)      171 non-null float64
열(천toe)         171 non-null float64
전력(천toe)        171 non-null float64
합계(천toe)        171 non-null float64
석탄류(천tCO2)      171 non-null float64
석유류(천tCO2)      171 non-null float64
가스류(천tCO2)      171 non-null float64
기타연료(천tCO2)     171 non-null float64
열(천tCO2)        171 non-null float64
전력(천tCO2)       171 non-null float64
합계(천tCO2)       171 non-null float64
dtypes: float64(16), int64(2), object(1)
memory usage: 25.5+ KB

에너지소비량 요인분해

[참고] 요인분해분석결과 단계별 컬럼정리

  • 에너지 : Q Si Ii QSiIi
  • CO2 : Qi Si Ii Mij Cij
  • 에너지 : Eit-Ei0 lnEit-lnEi0 ln(Qt/Q0) ln(Sit/Si0) ln(Iit/Ii0)
  • CO2 : Cijt-Cij0 lnCijt-lnCij0 ln(Qt/Q0) ln(Sit/Si0) ln(Iit/Ii0) ln(Mijt-Mij0)
  • 에너지 : 생산효과, 구조효과, 원단위개선효과 (Eact Estr Eins)
  • CO2 : 생산효과, 구조효과, 에너지원단위효과, 에너지믹스효과 (Cact Cstr Cint Cmix)
  • 에너지 : 소비량증감 Etot
  • CO2 : 배출량 증감 Ctot

분석용 데이터 사전처리

In [4]:
energy_df = df[['실적년도','감축업종코드','감축업종코드명','실질산출액(백만원)','실질부가가치액(백만원)','합계(천toe)',
               '석탄류(천toe)','석유류(천toe)','가스류(천toe)','기타연료(천toe)','열(천toe)','전력(천toe)']]
energy_df.columns =['year','code','name','out_value','add_value','energy','coal','petro','gas','others','heat','elec']
energy_df.head(3)
Out[4]:
year code name out_value add_value energy coal petro gas others heat elec
0 2011 1 광업 3.796199e+06 2.365625e+06 154.301071 35.379990 21.930911 0.078128 0.000 0.107800 96.804241
1 2011 2 음식료품 9.892679e+07 1.721800e+07 2096.328586 31.338552 356.569321 707.434415 9.210 157.935973 833.840325
2 2011 3 섬유 8.421345e+07 1.786797e+07 2196.294563 79.273859 195.799004 343.080913 11.827 348.709005 1217.604782

산업부문 전체 에너지 요인분해분석(실질산출액 기준)

In [7]:
def LMDI_out_energy(data, base_y, target_y):
    
    base  = data[data['year']== base_y].copy()      # 기준년도 데이터
    target  = data[data['year']== target_y].copy()  # 비교(실제)년도 데이터
    
    base['energy_b'] = base['energy']
    base['Q_b'] = sum(base['out_value'])/1000
    base['Si_b'] = (base['out_value']/1000)/ base['Q_b'] 
    base['li_b'] = base['energy']/(base['out_value']/1000)
    #base['QSili_b'] = base['Q_b']*base['Si_b']*base['li_b']
    base = base.drop(['coal','petro','gas','others','heat','elec','energy'], axis=1)
    
    target['energy_t'] = target['energy']
    target['Q_t'] = sum(target['out_value'])/1000
    target['Si_t'] = (target['out_value']/1000)/target['Q_t']
    target['li_t'] = target['energy']/(target['out_value']/1000)
    #target['QSili_t'] = target['Q_t']*target['Si_t']*target['li_t']
    target = target.drop(['name','out_value','add_value','coal','petro','gas','others','heat','elec','energy'], axis=1)
    
    result_df_temp = pd.merge(base, target, on='code')
    
    result_df = pd.DataFrame()
    result_df['Eit-Ei0'] = result_df_temp['energy_t']-result_df_temp['energy_b']
    result_df['lnEit-lnEi0'] = np.log(result_df_temp['energy_t'])-np.log(result_df_temp['energy_b'])
    result_df['ln(Qt/Q0)'] = np.log(result_df_temp['Q_t']/result_df_temp['Q_b'])
    result_df['ln(Sit/Si0)'] = np.log(result_df_temp['Si_t']/result_df_temp['Si_b'])
    result_df['ln(Iit/Ii0)'] = np.log(result_df_temp['li_t']/result_df_temp['li_b'])
    
    result_df['Eact'] = (result_df['Eit-Ei0']/result_df['lnEit-lnEi0'])*result_df['ln(Qt/Q0)']   # 생산효과
    result_df['Estr'] = (result_df['Eit-Ei0']/result_df['lnEit-lnEi0'])*result_df['ln(Sit/Si0)'] # 기준효과
    result_df['Eins'] = (result_df['Eit-Ei0']/result_df['lnEit-lnEi0'])*result_df['ln(Iit/Ii0)'] # 원단위개선효과
    result_df['Etot'] = result_df['Eact'] + result_df['Estr'] + result_df['Eins']                # 소비량증감

    print('< 산업전체 에너지 요인분해분석(실질산출액 기준),',base_y,'vs',target_y,'>\n')
    print('생산효과:',       np.round(sum(result_df['Eact']), 2))
    print('구조효과:',       np.round(sum(result_df['Estr']), 2))
    print('원단위효과:', np.round(sum(result_df['Eins']), 2))
    print('소비량증감:',     np.round(sum(result_df['Etot']), 2))
In [8]:
LMDI_out_energy(energy_df, 2015, 2018)

< 산업전체 에너지 요인분해분석(실질산출액 기준), 2015 vs 2018 >

생산효과: 7773.65
구조효과: 4095.65
원단위효과: 277.36
소비량증감: 12146.66

산업부문 전체 에너지 요인분해분석(실질부가가치액 기준)

In [13]:
def LMDI_add_energy(data, base_y, target_y):
    
    base  = data[data['year']== base_y].copy()      # 기준년도 데이터
    target  = data[data['year']== target_y].copy()  # 비교(실제)년도 데이터
    
    base['energy_b'] = base['energy']
    base['Q_b'] = sum(base['add_value'])/1000
    base['Si_b'] = (base['add_value']/1000)/ base['Q_b'] 
    base['li_b'] = base['energy']/(base['add_value']/1000)
    #base['QSili_b'] = base['Q_b']*base['Si_b']*base['li_b']
    base = base.drop(['coal','petro','gas','others','heat','elec','energy'], axis=1)
    
    target['energy_t'] = target['energy']
    target['Q_t'] = sum(target['add_value'])/1000
    target['Si_t'] = (target['add_value']/1000)/target['Q_t']
    target['li_t'] = target['energy']/(target['add_value']/1000)
    #target['QSili_t'] = target['Q_t']*target['Si_t']*target['li_t']
    target = target.drop(['name','add_value','add_value','coal','petro','gas','others','heat','elec','energy'], axis=1)
    
    result_df_temp = pd.merge(base, target, on='code')
    
    result_df = pd.DataFrame()
    result_df['Eit-Ei0'] = result_df_temp['energy_t']-result_df_temp['energy_b']
    result_df['lnEit-lnEi0'] = np.log(result_df_temp['energy_t'])-np.log(result_df_temp['energy_b'])
    result_df['ln(Qt/Q0)'] = np.log(result_df_temp['Q_t']/result_df_temp['Q_b'])
    result_df['ln(Sit/Si0)'] = np.log(result_df_temp['Si_t']/result_df_temp['Si_b'])
    result_df['ln(Iit/Ii0)'] = np.log(result_df_temp['li_t']/result_df_temp['li_b'])
    
    result_df['Eact'] = (result_df['Eit-Ei0']/result_df['lnEit-lnEi0'])*result_df['ln(Qt/Q0)']   # 생산효과
    result_df['Estr'] = (result_df['Eit-Ei0']/result_df['lnEit-lnEi0'])*result_df['ln(Sit/Si0)'] # 구조효과
    result_df['Eins'] = (result_df['Eit-Ei0']/result_df['lnEit-lnEi0'])*result_df['ln(Iit/Ii0)'] # 원단위개선효과
    result_df['Etot'] = result_df['Eact'] + result_df['Estr'] + result_df['Eins']                # 소비량증감

    print('< 산업전체 에너지 요인분해분석(실질부가가치액 기준),',base_y,'vs',target_y,'>\n')
    print('기준년도:', base_y, 'vs', '비교년도:', target_y) 
    print('생산효과:',       np.round(sum(result_df['Eact']), 2))
    print('구조효과:',       np.round(sum(result_df['Estr']), 2))
    print('원단위효과:', np.round(sum(result_df['Eins']), 2))
    print('소비량증감:',     np.round(sum(result_df['Etot']), 2))
In [14]:
LMDI_add_energy(energy_df, 2015, 2018)

< 산업전체 에너지 요인분해분석(실질부가가치액 기준), 2015 vs 2018 >

기준년도: 2015 vs 비교년도: 2018
생산효과: 11088.42
구조효과: -5325.43
원단위효과: 6383.67
소비량증감: 12146.66

산업부문 업종별 에너지 요인분해분석(실질산출액 기준)

In [23]:
def LMDI_sector_out_energy(data, base_y, target_y):
    
    base  = data[data['year']== base_y].copy()      # 기준년도 데이터
    target  = data[data['year']== target_y].copy()  # 비교(실제)년도 데이터
    
    base['energy_b'] = base['energy']
    base['Q_b'] = base['out_value']/1000
    base['li_b'] = base['energy']/(base['out_value']/1000)
    base = base.drop(['coal','petro','gas','others','heat','elec','energy'], axis=1)
    
    target['energy_t'] = target['energy']
    target['Q_t'] = target['out_value']/1000
    target['li_t'] = target['energy']/(target['out_value']/1000)
    target = target.drop(['name','out_value','add_value','coal','petro','gas','others','heat','elec','energy'], axis=1)
    
    result_df_temp = pd.merge(base, target, on='code')
    
    result_df = pd.DataFrame()
    result_df['Eit-Ei0'] = result_df_temp['energy_t']-result_df_temp['energy_b']
    result_df['lnEit-lnEi0'] = np.log(result_df_temp['energy_t'])-np.log(result_df_temp['energy_b'])
    result_df['ln(Qt/Q0)'] = np.log(result_df_temp['Q_t']/result_df_temp['Q_b'])
    result_df['ln(Iit/Ii0)'] = np.log(result_df_temp['li_t']/result_df_temp['li_b'])
    
    result_df_last = pd.DataFrame()
    result_df_last['업종명'] = result_df_temp['name']
    result_df_last['생산효과'] = (result_df['Eit-Ei0']/result_df['lnEit-lnEi0'])*result_df['ln(Qt/Q0)']     # 생산효과
    result_df_last['원단위효과'] = (result_df['Eit-Ei0']/result_df['lnEit-lnEi0'])*result_df['ln(Iit/Ii0)'] # 원단위효과
    result_df_last['소비량증감'] = result_df_last['생산효과'] + result_df_last['원단위효과']                # 소비량증감

    print('< 업종별 에너지 요인분해분석(실질산출액 기준),',base_y,'vs',target_y,'>\n')
    return result_df_last
In [24]:
LMDI_sector_out_energy(energy_df, 2015, 2018)

< 업종별 에너지 요인분해분석(실질산출액 기준), 2015 vs 2018 >
Out[24]:
업종명 생산효과 원단위효과 소비량증감
0 광업 -4.528632 8.629997 4.101365
1 음식료품 144.889982 284.782299 429.672281
2 섬유 -218.321876 172.420122 -45.901754
3 목재 26.952878 -102.123254 -75.170376
4 제지 32.672597 162.276369 194.948965
5 정유 5345.462260 -1265.844106 4079.618155
6 석유화학 2795.996919 526.064636 3322.061555
7 유리 30.401605 -28.727519 1.674085
8 요업 72.741891 -7.574957 65.166934
9 시멘트 682.625961 -1002.910000 -320.284039
10 철강 1552.493192 444.785088 1997.278280
11 비철금속 64.496919 169.968379 234.465298
12 기계 383.340807 390.038483 773.379290
13 반도체 1370.557377 -535.584167 834.973210
14 디스플레이 4.167798 53.524521 57.692319
15 전기전자 -91.284353 484.197200 392.912847
16 자동차 -110.275799 328.261365 217.985565
17 조선 -273.960812 192.007847 -81.952965
18 기타제조 60.869248 3.172065 64.041313

산업부문 업종별 에너지 요인분해분석(실질부가가치액 기준)

In [21]:
def LMDI_sector_add_energy(data, base_y, target_y):
    
    base  = data[data['year']== base_y].copy()      # 기준년도 데이터
    target  = data[data['year']== target_y].copy()  # 비교(실제)년도 데이터
    
    base['energy_b'] = base['energy']
    base['Q_b'] = base['add_value']/1000
    base['li_b'] = base['energy']/(base['add_value']/1000)
    base = base.drop(['coal','petro','gas','others','heat','elec','energy'], axis=1)
    
    target['energy_t'] = target['energy']
    target['Q_t'] = target['add_value']/1000
    target['li_t'] = target['energy']/(target['add_value']/1000)
    target = target.drop(['name','out_value','add_value','coal','petro','gas','others','heat','elec','energy'], axis=1)
    
    index_df_temp = pd.merge(base, target, on='code')
    
    index_df = pd.DataFrame()
    index_df['Eit-Ei0'] = index_df_temp['energy_t']-index_df_temp['energy_b']
    index_df['lnEit-lnEi0'] = np.log(index_df_temp['energy_t'])-np.log(index_df_temp['energy_b'])
    index_df['ln(Qt/Q0)'] = np.log(index_df_temp['Q_t']/index_df_temp['Q_b'])
    index_df['ln(Iit/Ii0)'] = np.log(index_df_temp['li_t']/index_df_temp['li_b'])
    
    result_df = pd.DataFrame()
    result_df['업종명'] = index_df_temp['name']
    result_df['생산효과'] = (index_df['Eit-Ei0']/index_df['lnEit-lnEi0'])*index_df['ln(Qt/Q0)']     # 생산효과
    result_df['원단위효과'] = (index_df['Eit-Ei0']/index_df['lnEit-lnEi0'])*index_df['ln(Iit/Ii0)'] # 원단위개선효과
    result_df['소비량증감'] = result_df['생산효과'] + result_df['원단위효과']                       # 소비량증감

    print('< 업종별 에너지 요인분해분석(실질부가가치액 기준),',base_y,'vs',target_y,'>\n')
    return result_df
In [22]:
LMDI_sector_add_energy(energy_df, 2015, 2018)

< 업종별 에너지 요인분해분석(실질부가가치액 기준), 2015 vs 2018 >
Out[22]:
업종명 생산효과 원단위효과 소비량증감
0 광업 -6.055879 10.157244 4.101365
1 음식료품 159.089228 270.583053 429.672281
2 섬유 -343.585710 297.683956 -45.901754
3 목재 21.356790 -96.527167 -75.170376
4 제지 122.501247 72.447718 194.948965
5 정유 809.354810 3270.263345 4079.618155
6 석유화학 2110.197783 1211.863772 3322.061555
7 유리 62.689808 -61.015723 1.674085
8 요업 86.629813 -21.462878 65.166934
9 시멘트 836.387996 -1156.672035 -320.284039
10 철강 740.035842 1257.242439 1997.278280
11 비철금속 -153.825671 388.290969 234.465298
12 기계 322.407497 450.971793 773.379290
13 반도체 1403.658308 -568.685098 834.973210
14 디스플레이 -87.389756 145.082075 57.692319
15 전기전자 -93.779580 486.692428 392.912847
16 자동차 -134.950565 352.936131 217.985565
17 조선 -126.695350 44.742385 -81.952965
18 기타제조 34.965906 29.075407 64.041313

이산화탄소배출량 요인분해

분석용 데이터 작업

In [27]:
energy_df = df[['실적년도','감축업종코드','감축업종코드명','실질산출액(백만원)','실질부가가치액(백만원)','합계(천toe)',
               '석탄류(천toe)','석유류(천toe)','가스류(천toe)','기타연료(천toe)','열(천toe)','전력(천toe)']]
energy_df.columns =['year','code','name','out_value','add_value','t_energy','coal','petro','gas','others','heat','elec']
energy_df_2 = energy_df.drop('t_energy', axis=1)
energy_df_2 = pd.melt(energy_df_2, id_vars=['year','code','name','out_value','add_value'])
energy_df_2 = energy_df_2.sort_values(['year','code'])
energy_df_2.columns =['year','code','name','out_value','add_value','energy_name','energy']
energy_df_2.head()
Out[27]:
year code name out_value add_value energy_name energy
0 2011 1 광업 3796199.0 2365625.0 coal 35.379990
171 2011 1 광업 3796199.0 2365625.0 petro 21.930911
342 2011 1 광업 3796199.0 2365625.0 gas 0.078128
513 2011 1 광업 3796199.0 2365625.0 others 0.000000
684 2011 1 광업 3796199.0 2365625.0 heat 0.107800
In [28]:
CO2_df_sub = df[['실적년도','감축업종코드','합계(천tCO2)','석탄류(천tCO2)','석유류(천tCO2)','가스류(천tCO2)','기타연료(천tCO2)','열(천tCO2)','전력(천tCO2)']]
CO2_df_sub.columns =['year','code','t_CO2','coal','petro','gas','others','heat','elec']
CO2_df_sub_2 = CO2_df_sub.drop('t_CO2', axis=1)
CO2_df_sub_2 = pd.melt(CO2_df_sub_2, id_vars=['year','code'])
CO2_df_sub_2 = CO2_df_sub_2.sort_values(['year','code'])
CO2_df_sub_2.columns =['year','code','energy_name','CO2']
CO2_df_sub_2.head(10)
Out[28]:
year code energy_name CO2
0 2011 1 coal 139.436559
171 2011 1 petro 70.335699
342 2011 1 gas 0.168250
513 2011 1 others 0.000000
684 2011 1 heat 0.395267
855 2011 1 elec 509.811080
1 2011 2 coal 117.007696
172 2011 2 petro 987.798086
343 2011 2 gas 1510.081460
514 2011 2 others 33.770000
In [29]:
CO2_df = pd.merge(energy_df_2, CO2_df_sub_2, on=['year','code','energy_name']) # 업종별 에너지사용량, 온실가스배출량
CO2_df.head()
Out[29]:
year code name out_value add_value energy_name energy CO2
0 2011 1 광업 3796199.0 2365625.0 coal 35.379990 139.436559
1 2011 1 광업 3796199.0 2365625.0 petro 21.930911 70.335699
2 2011 1 광업 3796199.0 2365625.0 gas 0.078128 0.168250
3 2011 1 광업 3796199.0 2365625.0 others 0.000000 0.000000
4 2011 1 광업 3796199.0 2365625.0 heat 0.107800 0.395267
In [30]:
energy_df_s = energy_df[['year','code','t_energy']] # 업종별 총 에너지사용량 출력 
CO2_df_sub_s = CO2_df_sub[['year','code','t_CO2']]  # 업종별 총 CO2배출량 출력
In [31]:
CO2_df = pd.merge(CO2_df, energy_df_s, on=['year','code'])
CO2_df = pd.merge(CO2_df, CO2_df_sub_s, on=['year','code'])
CO2_df.head()
Out[31]:
year code name out_value add_value energy_name energy CO2 t_energy t_CO2
0 2011 1 광업 3796199.0 2365625.0 coal 35.379990 139.436559 154.301071 720.146855
1 2011 1 광업 3796199.0 2365625.0 petro 21.930911 70.335699 154.301071 720.146855
2 2011 1 광업 3796199.0 2365625.0 gas 0.078128 0.168250 154.301071 720.146855
3 2011 1 광업 3796199.0 2365625.0 others 0.000000 0.000000 154.301071 720.146855
4 2011 1 광업 3796199.0 2365625.0 heat 0.107800 0.395267 154.301071 720.146855

산업부문 전체 이산화탄소 요인분해분석(실질산출액 기준)

In [34]:
def LMDI_out_CO2(data, base_y, target_y):
    
    base  = data[data['year']== base_y].copy()      # 기준년도 데이터
    target  = data[data['year']== target_y].copy()  # 비교(실제)년도 데이터
    
    base['CO2_b'] = base['CO2']
    base['Qi_b'] = (sum(base['out_value'])/6)/1000
    base['Si_b'] = (base['out_value']/1000)/ base['Qi_b'] 
    base['li_b'] = base['t_energy']/(base['out_value']/1000)
    base['Mij_b'] = base['CO2']/base['t_energy']
    base['Cij_b'] = base['Qi_b']*base['Si_b']*base['li_b']*base['Mij_b']
    
    target['CO2_t'] =target['CO2']
    target['Qi_t'] = (sum(target['out_value'])/6)/1000
    target['Si_t'] = (target['out_value']/1000)/ target['Qi_t'] 
    target['li_t'] = target['t_energy']/(target['out_value']/1000)
    target['Mij_t'] = target['CO2']/target['t_energy']
    target['Cij_t'] = target['Qi_t']*target['Si_t']*target['li_t']*target['Mij_t']

    result_df_temp = pd.merge(base, target, on=['code', 'energy_name'])
    
    result_df = pd.DataFrame()
    result_df['Cijt-Cij0'] = result_df_temp['CO2_t']-result_df_temp['CO2_b']
    result_df['lnCijt-lnCij0'] = np.log(result_df_temp['CO2_t'])-np.log(result_df_temp['CO2_b'])
    result_df['ln(Qt/Q0)'] = np.log(result_df_temp['Qi_t']/result_df_temp['Qi_b'])
    result_df['ln(Sit/Si0)'] = np.log(result_df_temp['Si_t']/result_df_temp['Si_b'])
    result_df['ln(Iit/Ii0)'] = np.log(result_df_temp['li_t']/result_df_temp['li_b'])
    result_df['ln(Mijt-Mij0)'] = np.log(result_df_temp['Mij_t']/result_df_temp['Mij_b'])
    
    result_df['Cact'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Qt/Q0)']     # 생산효과
    result_df['Cstr'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Sit/Si0)']   # 구조효과
    result_df['Cint'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Iit/Ii0)']   # 원단위개선효과
    result_df['Cmix'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Mijt-Mij0)'] # 에너지믹스효과
    result_df['Ctot'] = result_df['Cact'] + result_df['Cstr'] + result_df['Cint'] + result_df['Cmix']   # 배출량증감

    result_df = result_df.fillna(0)
   #return result_df

    
    print('< 산업전체 CO2 요인분해분석(실질산출액 기준),',base_y,'vs',target_y,'>\n')
    print('생산효과:',         np.round(sum(result_df['Cact']), 2))
    print('구조효과:',         np.round(sum(result_df['Cstr']), 2))
    print('원단위효과:',       np.round(sum(result_df['Cint']), 2))
    print('에너지믹스효과:',   np.round(sum(result_df['Cmix']), 2))
    print('배출량증감:',       np.round(sum(result_df['Ctot']), 2))
In [35]:
LMDI_out_CO2(CO2_df, 2015, 2018)

< 산업전체 CO2 요인분해분석(실질산출액 기준), 2015 vs 2018 >

생산효과: 20258.11
구조효과: 6034.36
원단위효과: 3938.94
에너지믹스효과: -368.19
배출량증감: 29863.22

C:\ProgramData\Anaconda3\lib\site-packages\ipykernel_launcher.py:24: RuntimeWarning: divide by zero encountered in log
C:\ProgramData\Anaconda3\lib\site-packages\ipykernel_launcher.py:28: RuntimeWarning: divide by zero encountered in log

산업부문 전체 이산화탄소 요인분해분석(실질부가가치액 기준)

In [38]:
def LMDI_add_CO2(data, base_y, target_y):
    
    base  = data[data['year']== base_y].copy()      # 기준년도 데이터
    target  = data[data['year']== target_y].copy()  # 비교(실제)년도 데이터
    
    base['CO2_b'] = base['CO2']
    base['Qi_b'] = (sum(base['add_value'])/6)/1000
    base['Si_b'] = (base['add_value']/1000)/ base['Qi_b'] 
    base['li_b'] = base['t_energy']/(base['add_value']/1000)
    base['Mij_b'] = base['CO2']/base['t_energy']
    base['Cij_b'] = base['Qi_b']*base['Si_b']*base['li_b']*base['Mij_b']
    
    target['CO2_t'] =target['CO2']
    target['Qi_t'] = (sum(target['add_value'])/6)/1000
    target['Si_t'] = (target['add_value']/1000)/ target['Qi_t'] 
    target['li_t'] = target['t_energy']/(target['add_value']/1000)
    target['Mij_t'] = target['CO2']/target['t_energy']
    target['Cij_t'] = target['Qi_t']*target['Si_t']*target['li_t']*target['Mij_t']

    result_df_temp = pd.merge(base, target, on=['code', 'energy_name'])
    
    result_df = pd.DataFrame()
    result_df['Cijt-Cij0'] = result_df_temp['CO2_t']-result_df_temp['CO2_b']
    result_df['lnCijt-lnCij0'] = np.log(result_df_temp['CO2_t'])-np.log(result_df_temp['CO2_b'])
    result_df['ln(Qt/Q0)'] = np.log(result_df_temp['Qi_t']/result_df_temp['Qi_b'])
    result_df['ln(Sit/Si0)'] = np.log(result_df_temp['Si_t']/result_df_temp['Si_b'])
    result_df['ln(Iit/Ii0)'] = np.log(result_df_temp['li_t']/result_df_temp['li_b'])
    result_df['ln(Mijt-Mij0)'] = np.log(result_df_temp['Mij_t']/result_df_temp['Mij_b'])
    
    result_df['Cact'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Qt/Q0)']     # 생산효과
    result_df['Cstr'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Sit/Si0)']   # 구조효과
    result_df['Cint'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Iit/Ii0)']   # 원단위개선효과
    result_df['Cmix'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Mijt-Mij0)'] # 에너지믹스효과
    result_df['Ctot'] = result_df['Cact'] + result_df['Cstr'] + result_df['Cint'] + result_df['Cmix']   # 배출량증감

    result_df = result_df.fillna(0)
   #return result_df

    
    print('< 산업전체 CO2 요인분해분석(실질부가가치액 기준),',base_y,'vs',target_y,'>\n')
    print('생산효과:',         np.round(sum(result_df['Cact']), 2))
    print('구조효과:',         np.round(sum(result_df['Cstr']), 2))
    print('원단위효과:',       np.round(sum(result_df['Cint']), 2))
    print('에너지믹스효과:',   np.round(sum(result_df['Cmix']), 2))
    print('배출량증감:',       np.round(sum(result_df['Ctot']), 2))
In [39]:
LMDI_add_CO2(CO2_df, 2015, 2018)

< 산업전체 CO2 요인분해분석(실질부가가치액 기준), 2015 vs 2018 >

생산효과: 28896.4
구조효과: -12154.17
원단위효과: 13489.17
에너지믹스효과: -368.19
배출량증감: 29863.22

C:\ProgramData\Anaconda3\lib\site-packages\ipykernel_launcher.py:24: RuntimeWarning: divide by zero encountered in log
C:\ProgramData\Anaconda3\lib\site-packages\ipykernel_launcher.py:28: RuntimeWarning: divide by zero encountered in log

산업부문 업종별 이산화탄소 요인분해분석(실질산출액 기준)

In [40]:
CO2_df.head()
Out[40]:
year code name out_value add_value energy_name energy CO2 t_energy t_CO2
0 2011 1 광업 3796199.0 2365625.0 coal 35.379990 139.436559 154.301071 720.146855
1 2011 1 광업 3796199.0 2365625.0 petro 21.930911 70.335699 154.301071 720.146855
2 2011 1 광업 3796199.0 2365625.0 gas 0.078128 0.168250 154.301071 720.146855
3 2011 1 광업 3796199.0 2365625.0 others 0.000000 0.000000 154.301071 720.146855
4 2011 1 광업 3796199.0 2365625.0 heat 0.107800 0.395267 154.301071 720.146855
In [64]:
def LMDI_sector_out_CO2(data, base_y, target_y):
    
    base  = data[data['year']== base_y].copy()      # 기준년도 데이터
    target  = data[data['year']== target_y].copy()  # 비교(실제)년도 데이터
    
    base['CO2_b'] = base['CO2']
    base['Qi_b'] = base['out_value']/1000
    base['li_b'] = base['t_energy']/(base['out_value']/1000)
    base['Mij_b'] = base['CO2']/base['t_energy']
    base['Cij_b'] = base['Qi_b']*base['li_b']*base['Mij_b']
    
    target['CO2_t'] =target['CO2']
    target['Qi_t'] = target['out_value']/1000
    target['li_t'] = target['t_energy']/(target['out_value']/1000)
    target['Mij_t'] = target['CO2']/target['t_energy']
    target['Cij_t'] = target['Qi_t']*target['li_t']*target['Mij_t']

    result_df_temp = pd.merge(base, target, on=['code', 'name','energy_name'])
    
    result_df = pd.DataFrame()
    result_df['code'] = result_df_temp['code']
    result_df['Cijt-Cij0'] = result_df_temp['CO2_t']-result_df_temp['CO2_b']
    result_df['lnCijt-lnCij0'] = np.log(result_df_temp['CO2_t'])-np.log(result_df_temp['CO2_b'])
    result_df['ln(Qt/Q0)'] = np.log(result_df_temp['Qi_t']/result_df_temp['Qi_b'])
    result_df['ln(Iit/Ii0)'] = np.log(result_df_temp['li_t']/result_df_temp['li_b'])
    result_df['ln(Mijt-Mij0)'] = np.log(result_df_temp['Mij_t']/result_df_temp['Mij_b'])
    
    result_df['Cact'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Qt/Q0)']     # 생산효과
    result_df['Cint'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Iit/Ii0)']   # 원단위개선효과
    result_df['Cmix'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Mijt-Mij0)'] # 에너지믹스효과
    result_df['Ctot'] = result_df['Cact'] + result_df['Cint'] + result_df['Cmix']   # 배출량증감

    result_df = result_df.fillna(0)
    
    index_df = result_df_temp[['code', 'name']].drop_duplicates()      # 업종명 열 생성
    result_df_sub = result_df[['code','Cact','Cint','Cmix','Ctot']]    # 분석효과 열만 추출
    result_df_sub = result_df_sub.groupby(['code']).sum()
    result_df_last = pd.merge(index_df, result_df_sub, on='code')
    result_df_last = result_df_last.drop(['code'], axis=1)
    result_df_last.columns = ['업종명','생산효과','원단위효과','에너지믹스효과','배출량증감']
    return result_df_last
In [67]:
LMDI_sector_out_CO2(CO2_df, 2015, 2018)

C:\ProgramData\Anaconda3\lib\site-packages\ipykernel_launcher.py:23: RuntimeWarning: divide by zero encountered in log
C:\ProgramData\Anaconda3\lib\site-packages\ipykernel_launcher.py:26: RuntimeWarning: divide by zero encountered in log
Out[67]:
업종명 생산효과 원단위효과 에너지믹스효과 배출량증감
0 광업 -20.997003 40.012980 3.691018 22.706996
1 음식료품 547.010169 1075.152412 282.113517 1904.276098
2 섬유 -867.821456 685.363665 -374.396736 -556.854528
3 목재 105.361608 -399.210442 -73.989764 -367.838598
4 제지 127.737867 634.441072 -575.611672 186.567267
5 정유 5631.134442 -1333.493343 -2850.874141 1446.766957
6 석유화학 4960.721630 933.355899 1850.410253 7744.487783
7 유리 111.406906 -105.272207 28.490458 34.625156
8 요업 245.477135 -25.562694 -92.571261 127.343180
9 시멘트 2617.985306 -3846.328435 35.659497 -1192.683631
10 철강 6032.435067 1728.276282 2088.391575 9849.102924
11 비철금속 261.505645 689.144399 -235.118098 715.531946
12 기계 1738.001707 1768.367824 389.855124 3896.224654
13 반도체 6569.474132 -2567.208341 -190.456419 3811.809371
14 디스플레이 20.455642 262.699506 -136.217424 146.937724
15 전기전자 -417.722623 2215.715151 -126.648448 1671.344080
16 자동차 -441.454048 1314.089845 -380.996845 491.638952
17 조선 -1218.867170 854.253786 -74.217840 -438.831224
18 기타제조 290.621156 15.145074 64.294566 370.060796

산업부문 업종별 이산화탄소 요인분해분석(실질부가가치액 기준)

In [68]:
def LMDI_sector_add_CO2(data, base_y, target_y):
    
    base  = data[data['year']== base_y].copy()      # 기준년도 데이터
    target  = data[data['year']== target_y].copy()  # 비교(실제)년도 데이터
    
    base['CO2_b'] = base['CO2']
    base['Qi_b'] = base['add_value']/1000
    base['li_b'] = base['t_energy']/(base['add_value']/1000)
    base['Mij_b'] = base['CO2']/base['t_energy']
    base['Cij_b'] = base['Qi_b']*base['li_b']*base['Mij_b']
    
    target['CO2_t'] =target['CO2']
    target['Qi_t'] = target['add_value']/1000
    target['li_t'] = target['t_energy']/(target['add_value']/1000)
    target['Mij_t'] = target['CO2']/target['t_energy']
    target['Cij_t'] = target['Qi_t']*target['li_t']*target['Mij_t']

    result_df_temp = pd.merge(base, target, on=['code', 'name','energy_name'])
    
    result_df = pd.DataFrame()
    result_df['code'] = result_df_temp['code']
    result_df['Cijt-Cij0'] = result_df_temp['CO2_t']-result_df_temp['CO2_b']
    result_df['lnCijt-lnCij0'] = np.log(result_df_temp['CO2_t'])-np.log(result_df_temp['CO2_b'])
    result_df['ln(Qt/Q0)'] = np.log(result_df_temp['Qi_t']/result_df_temp['Qi_b'])
    result_df['ln(Iit/Ii0)'] = np.log(result_df_temp['li_t']/result_df_temp['li_b'])
    result_df['ln(Mijt-Mij0)'] = np.log(result_df_temp['Mij_t']/result_df_temp['Mij_b'])
    
    result_df['Cact'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Qt/Q0)']     # 생산효과
    result_df['Cint'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Iit/Ii0)']   # 원단위개선효과
    result_df['Cmix'] = (result_df['Cijt-Cij0']/result_df['lnCijt-lnCij0'])*result_df['ln(Mijt-Mij0)'] # 에너지믹스효과
    result_df['Ctot'] = result_df['Cact'] + result_df['Cint'] + result_df['Cmix']   # 배출량증감

    result_df = result_df.fillna(0)
    
    index_df = result_df_temp[['code', 'name']].drop_duplicates()      # 업종명 열 생성
    result_df_sub = result_df[['code','Cact','Cint','Cmix','Ctot']]    # 분석효과 열만 추출
    result_df_sub = result_df_sub.groupby(['code']).sum()
    result_df_last = pd.merge(index_df, result_df_sub, on='code')
    result_df_last = result_df_last.drop(['code'], axis=1)
    result_df_last.columns = ['업종명','생산효과','원단위효과','에너지믹스효과','배출량증감']
    return result_df_last
In [69]:
LMDI_sector_add_CO2(CO2_df, 2015, 2018)

C:\ProgramData\Anaconda3\lib\site-packages\ipykernel_launcher.py:23: RuntimeWarning: divide by zero encountered in log
C:\ProgramData\Anaconda3\lib\site-packages\ipykernel_launcher.py:26: RuntimeWarning: divide by zero encountered in log
Out[69]:
업종명 생산효과 원단위효과 에너지믹스효과 배출량증감
0 광업 -28.078085 47.094063 3.691018 22.706996
1 음식료품 600.617270 1021.545310 282.113517 1904.276098
2 섬유 -1365.740606 1183.282815 -374.396736 -556.854528
3 목재 83.485920 -377.334754 -73.989764 -367.838598
4 제지 478.934939 283.244000 -575.611672 186.567267
5 정유 852.608348 3445.032751 -2850.874141 1446.766957
6 석유화학 3743.961130 2150.116399 1850.410253 7744.487783
7 유리 229.727265 -223.592567 28.490458 34.625156
8 요업 292.343763 -72.429322 -92.571261 127.343180
9 시멘트 3207.688559 -4436.031688 35.659497 -1192.683631
10 철강 2875.515451 4885.195898 2088.391575 9849.102924
11 비철금속 -623.693068 1574.343112 -235.118098 715.531946
12 기계 1461.740494 2044.629037 389.855124 3896.224654
13 반도체 6728.136374 -2725.870583 -190.456419 3811.809371
14 디스플레이 -428.910816 712.065963 -136.217424 146.937724
15 전기전자 -429.140931 2227.133459 -126.648448 1671.344080
16 자동차 -540.231615 1412.867412 -380.996845 491.638952
17 조선 -563.674787 199.061403 -74.217840 -438.831224
18 기타제조 166.945253 138.820977 64.294566 370.060796
In [ ]: