Loading Libraries

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt 
%matplotlib inline
import matplotlib as mpl
mpl.rcParams['figure.figsize'] = (10,5)
import seaborn as sns
import statsmodels.api as sm
import statsmodels.graphics.api as smg

Loading Data

co2 = pd.read_csv('co2.csv')
tmpr = pd.read_csv('temperature.csv')

Loaded 2 dataset from “datahub.io”:

1. Global CO2 Emissions from Fossil Fuels since 1751 
2. Global Temperature Time Series

---

Global CO2 Emissions

Global CO2 Emissions from fossil-fuels annually since 1751 till 2014.

source: https://datahub.io/core/co2-fossil-global

Global Temperature Time Series

Global Temperature Time Series. Data are included from the GISS Surface Temperature (GISTEMP) analysis and the global component of Climate at a Glance (GCAG). Two datasets are provided: 1) global monthly mean and 2) annual mean temperature anomalies in degrees Celsius from 1880 to the present.

source: https://datahub.io/core/global-temp

co2.head()

	Year	Total	Gas Fuel	Liquid Fuel	Solid Fuel	Cement	Gas Flaring	Per Capita
0	1751	3	0	0	3	0	0	NaN
1	1752	3	0	0	3	0	0	NaN
2	1753	3	0	0	3	0	0	NaN
3	1754	3	0	0	3	0	0	NaN
4	1755	3	0	0	3	0	0	NaN

co2.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 260 entries, 0 to 259
Data columns (total 8 columns):
 #   Column       Non-Null Count  Dtype  
---  ------       --------------  -----  
 0   Year         260 non-null    int64  
 1   Total        260 non-null    int64  
 2   Gas Fuel     260 non-null    int64  
 3   Liquid Fuel  260 non-null    int64  
 4   Solid Fuel   260 non-null    int64  
 5   Cement       260 non-null    int64  
 6   Gas Flaring  260 non-null    int64  
 7   Per Capita   61 non-null     float64
dtypes: float64(1), int64(7)
memory usage: 16.4 KB

co2.describe()

	Year	Total	Gas Fuel	Liquid Fuel	Solid Fuel	Cement	Gas Flaring	Per Capita
count	260.000000	260.000000	260.00000	260.000000	260.000000	260.000000	260.000000	61.000000
mean	1880.500000	1402.788462	185.20000	495.819231	674.569231	34.161538	13.065385	1.054754
std	75.199734	2253.098527	396.58556	934.308074	868.368580	78.899604	26.311315	0.178630
min	1751.000000	3.000000	0.00000	0.000000	3.000000	0.000000	0.000000	0.640000
25%	1815.750000	12.750000	0.00000	0.000000	12.750000	0.000000	0.000000	0.940000
50%	1880.500000	239.500000	0.00000	3.000000	236.000000	0.000000	0.000000	1.120000
75%	1945.250000	1385.000000	59.50000	279.250000	1023.500000	12.000000	0.000000	1.170000
max	2010.000000	9167.000000	1702.00000	3122.000000	3842.000000	450.000000	110.000000	1.330000

tmpr.head()

	Source	Year	Mean
0	GCAG	2016	0.9363
1	GISTEMP	2016	0.9900
2	GCAG	2015	0.8998
3	GISTEMP	2015	0.8700
4	GCAG	2014	0.7408

tmpr.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 274 entries, 0 to 273
Data columns (total 3 columns):
 #   Column  Non-Null Count  Dtype  
---  ------  --------------  -----  
 0   Source  274 non-null    object 
 1   Year    274 non-null    int64  
 2   Mean    274 non-null    float64
dtypes: float64(1), int64(1), object(1)
memory usage: 6.5+ KB

tmpr.describe()

	Year	Mean
count	274.000000	274.000000
mean	1948.000000	0.036588
std	39.619805	0.320069
min	1880.000000	-0.470000
25%	1914.000000	-0.204350
50%	1948.000000	-0.053400
75%	1982.000000	0.229025
max	2016.000000	0.990000

Data Cleansing

We have to filter ‘Global CO2 Emissions’ for years after 1880 because ‘Global Temperature’ data is after 1880.

# Filtering dataset to match other dataset
co2 = co2[co2['Year'] >= 1880]
co2

	Year	Total	Gas Fuel	Liquid Fuel	Solid Fuel	Cement	Gas Flaring	Per Capita
129	1880	236	0	3	233	0	0	NaN
130	1881	243	0	4	239	0	0	NaN
131	1882	256	0	4	252	0	0	NaN
132	1883	272	0	3	269	0	0	NaN
133	1884	275	0	4	271	0	0	NaN
...	...	...	...	...	...	...	...	...
255	2006	8370	1525	3089	3339	356	61	1.27
256	2007	8566	1572	3081	3464	382	68	1.28
257	2008	8783	1631	3122	3571	388	71	1.30
258	2009	8740	1585	3056	3620	413	66	1.28
259	2010	9167	1702	3114	3842	450	59	1.33

131 rows × 8 columns

We have to filter ‘Global Temperature’ for years before 2010 because ‘Global CO2 Emissions’ data is until 2010.

# Filtering dataset to match other dataset
tmpr = tmpr[tmpr['Year'] <= 2010]
tmpr

	Source	Year	Mean
12	GCAG	2010	0.7014
13	GISTEMP	2010	0.7100
14	GCAG	2009	0.6367
15	GISTEMP	2009	0.6400
16	GCAG	2008	0.5419
...	...	...	...
269	GISTEMP	1882	-0.1000
270	GCAG	1881	-0.0628
271	GISTEMP	1881	-0.1200
272	GCAG	1880	-0.1148
273	GISTEMP	1880	-0.2000

262 rows × 3 columns

For ‘Global CO2 Emissions’ keep only ‘Total’ column and drop others, set the index to ‘Year’ Column and sort data by index(Year).

co2 = co2.drop(['Gas Fuel', 'Liquid Fuel', 'Solid Fuel', 'Cement', 'Gas Flaring', 'Per Capita'], axis=1).set_index('Year').sort_index()

---

For ‘Global CO2 Emissions’ dataset, separating two sources (GCAG, GISTEMP), then drop the source and set the index on ‘Year’ column and sort data by index (Year)

GCAG = tmpr[(tmpr['Source'] == 'GCAG')].set_index('Year').drop(['Source'], axis=1).sort_index()

GISTEMP = tmpr[(tmpr['Source'] == 'GISTEMP')].set_index('Year').drop(['Source'], axis=1).sort_index()

Insights

Set a color for each data set and plot data to visually check our datasets.

• ‘Global CO2 Emissions’ = ‘#5F9EA0’
• ‘Global Temperature GCAG’ = ‘#A52A2A’
• ‘Global Temperature GIS_c’ = ‘#8A2BE2’

#set a color
co2_c = '#5F9EA0' #Cadet blue
GCAG_c = '#A52A2A' #Brown
GIS_c = '#8A2BE2' #Blue violet
fig, ax1 = plt.subplots()
# Add CO2 dataset on second axis
ax2 = ax1.twinx()
ax1.plot(GCAG, color= GCAG_c, label='GCAG')
ax1.plot(GISTEMP, color = GIS_c, label='GISTEMP')
ax2.plot(co2, color = co2_c, label='CO2')
plt.legend()
plt.show()

There are some differences between GCAG, and GISTEMP data, comparing two datasets’s histogram.

ax1 = plt.hist(GCAG, 10, alpha = 0.5, color = GCAG_c, label='GCAG')
ax2 = plt.hist(GISTEMP, 10, alpha = 0.5, color = GIS_c, label='GISTEMP')
plt.legend()
plt.show()

Plot a scatter plot for each dataset (GCAG, GISTEMP) with CO2.

plt.subplot(1,2,1)
plt.scatter(GCAG, co2, alpha=0.5, s = 100, color = GCAG_c)
plt.title('GCAG')
plt.ylabel('CO2')
plt.subplot(1,2,2)
plt.scatter(GISTEMP, co2, alpha=0.5, s = 100, color = GIS_c)
plt.ylabel('CO2')
plt.title('GISTEMP')
plt.show()

Visually we can see a correlation between our datasets, but know we check the numbers

GCAG_corr = pd.concat([GCAG, co2], axis=1).corr()
GCAG_corr

	Mean	Total
Mean	1.000000	0.888284
Total	0.888284	1.000000

GISTEMP_corr = pd.concat([GISTEMP, co2], axis=1).corr()
GISTEMP_corr

	Mean	Total
Mean	1.000000	0.897529
Total	0.897529	1.000000

We can the there is a strong relationship between ‘Global CO2 Emissions’ & ‘Global Temperature’.

• GCAG = 0.888284
• GISTEMP = 0.897529

Linear Regression Models

GCAG_model = sm.OLS(GCAG, co2).fit()
GCAG_model.summary2()

Model:	OLS	Adj. R-squared (uncentered):	0.412
Dependent Variable:	Mean	AIC:	-30.0077
Date:	2021-02-11 16:06	BIC:	-27.1325
No. Observations:	131	Log-Likelihood:	16.004
Df Model:	1	F-statistic:	92.76
Df Residuals:	130	Prob (F-statistic):	6.70e-17
R-squared (uncentered):	0.416	Scale:	0.046211

	Coef.	Std.Err.	t	P>|t|	[0.025	0.975]
Total	0.0000	0.0000	9.6314	0.0000	0.0000	0.0001

Omnibus:	3.886	Durbin-Watson:	0.226
Prob(Omnibus):	0.143	Jarque-Bera (JB):	3.604
Skew:	0.337	Prob(JB):	0.165
Kurtosis:	2.546	Condition No.:	1

GISTEMP_model = sm.OLS(GISTEMP, co2).fit()
GISTEMP_model.summary2()

Model:	OLS	Adj. R-squared (uncentered):	0.338
Dependent Variable:	Mean	AIC:	-3.7803
Date:	2021-02-11 16:06	BIC:	-0.9051
No. Observations:	131	Log-Likelihood:	2.8901
Df Model:	1	F-statistic:	67.87
Df Residuals:	130	Prob (F-statistic):	1.63e-13
R-squared (uncentered):	0.343	Scale:	0.056453

	Coef.	Std.Err.	t	P>|t|	[0.025	0.975]
Total	0.0000	0.0000	8.2383	0.0000	0.0000	0.0001

Omnibus:	6.004	Durbin-Watson:	0.210
Prob(Omnibus):	0.050	Jarque-Bera (JB):	6.112
Skew:	0.500	Prob(JB):	0.047
Kurtosis:	2.655	Condition No.:	1

fig, ax = plt.subplots()
fig = sm.graphics.plot_fit(GCAG_model, 0, ax=ax)
ax.set_ylabel("GCAG")
ax.set_xlabel("CO2")
ax.set_title("Linear Regression")

Text(0.5, 1.0, 'Linear Regression')

fig, ax = plt.subplots()
fig = sm.graphics.plot_fit(GISTEMP_model, 0, ax=ax)
ax.set_ylabel("GCAG")
ax.set_xlabel("CO2")
ax.set_title("Linear Regression")

Text(0.5, 1.0, 'Linear Regression')

Using Seaborn 'regplot'

sns.regplot(x="Total", y="Mean", data=pd.concat([GCAG, co2], axis=1), scatter_kws={"color": "brown"}, line_kws={"color": "gray"});

sns.regplot(x="Total", y="Mean", data=pd.concat([GISTEMP, co2], axis=1), scatter_kws={"color": "blueviolet"}, line_kws={"color": "gray"}, label = 'GISTEMP');

Predict 2011-2021

pyear = np.arange(2011, 2021)
pd.DataFrame({'Year': pyear, 'GCAG': GCAG_model.predict(pyear), 'GISTEMP': GISTEMP_model.predict(pyear)})

	Year	GCAG	GISTEMP
0	2011	0.097440	0.092121
1	2012	0.097488	0.092166
2	2013	0.097536	0.092212
3	2014	0.097585	0.092258
4	2015	0.097633	0.092304
5	2016	0.097682	0.092350
6	2017	0.097730	0.092395
7	2018	0.097779	0.092441
8	2019	0.097827	0.092487
9	2020	0.097876	0.092533