IA mathematics

Title of exploration: what is the correlation between carbon dioxide

emission and total energy consumption?

Page number: 20 pages

1
Table of Contents
Introduction and rationale..............................................................................................................................3
Personal engagement...............................................................................................................................4
Aim..........................................................................................................................................................4
Methodology..................................................................................................................................................5
Data collection...............................................................................................................................................5
Exploration.....................................................................................................................................................9
Histograms..............................................................................................................................................9
Scatter plot.............................................................................................................................................11
Correlation analysis...............................................................................................................................13
Pearson correlation:...........................................................................................................................14
Linear regression...................................................................................................................................17
Conclusion and evaluation...........................................................................................................................21
Bibliography................................................................................................................................................23
Appendix......................................................................................................................................................24

2
Introduction and rationale
CO2 (carbon dioxide) emission is the release of carbon dioxide gas into the atmosphere,

which can be from different sources, measured in (billion metric tons) which is called a gigaton

(Gt). The emissions are a major contributor to global warming and climate change, as CO 2 is a

greenhouse gas that can trap heat in the earth’s atmosphere. Energy production, with all its types,

accounts for 72% of all CO2 emissions. So, it can be regarded as the main source of carbon

dioxide production. The demand for energy increases annually, which could cause a significant

issue in climate change.

Total energy consumption refers to the total amount of energy used within a time frame in

a specific region, which is measured in (trillion-watt hours) which is also called terawatt-hours

(TWh). The energy consumption of a regions depends on its economical state, population size,

industrialization, and other factors. The largest source of energy in the world is oil, making up to

32% of the world energy use in 2023. Fossil fuels are 81% of the world energy use.

4% 2% 2%

4% 6%
32%

23%

26%

Oil Coal Natural gas

Nuclear Hydro Wind
Solar Biomass and other renewables

Figure 1: percentage of world energy use by each source in pie chart

3
 Climate change is the long-term shifts in temperatures and weather patterns. More

frequent and intense droughts, storms, heat waves, rising sea levels, melting of glaciers and

warming are caused by the increase in climate change over the recent years. This has a direct

effect on animals, habitats, communities, agriculture, and therefore, us. This is why climate

change is an important issue that needs to be solved as fast as possible to reduce the extent of the

damage.

Personal engagement
As a citizen of this earth, I am obligated to take action in any way I can. I have witnessed

firsthand the consequences of climate change, which is primarily caused by CO2 emissions, while

visiting Alexandria, where coastal erosion is becoming more severe and sea is intruding further

inland. Not only that, but as an Egyptian citizen I am extremely worried about the Nile’s state

and water flow, which is affecting water distribution and, therefore, agriculture. Moreover,

temperatures in Egypt, especially during the summer, has become unbearable with intense heat

waves, unlike what it used to be like. Investigating a major factor that contributes to CO2

emission, total energy consumption, can help find a solution that can contribute, even if just a

little, to climate change.

Aim
The aim of the research is to explore the correlation between CO2 emissions and total

energy consumption. By using statistical models, I will investigate the correlation between the 2

variables. Understanding the relationship can provide helpful insights into how energy

consumption patterns affect the carbon footprint, which will shed light into a potential cause of

the issue. This will help guide the efforts to reduce emissions by improving energy efficiency.

4
Methodology
The analysis explores the correlation between global carbon dioxide emissions and total

energy consumption using statistical methods. Data for carbon dioxide emissions (in gigaton)

and energy consumption (in terawatt hours) were collected from reliable sources and aligned

chronologically. Descriptive statistics, incorporated in histograms, were calculated to understand

the data distribution. To provide visualization on the data, I will create a scatter plot to assess the

trendline from which I will calculate the equation of the line in order to test the linearity if found.

Pearson’s correlation coefficient will be used to measure the linear relationship between the two

variables and their strength. Spearman’s rank correlation will also be used to further assess

monotonic trends, if the Pearson’s correlation coefficient is moderate or less. However, if

Pearson’s correlation coefficient was strong, a linear regression analysis will be carried out to

quantify the relationship between the two variables and predict future values.

Data collection
The data for this investigation was sourced from historical graphs and datasets, mainly

from “our world in data”1 and “statista”2. These data points were extracted and organized from

the downloadable excel files of emissions trends spanning from 1940 to 2023, as well as

corresponding energy consumption trends. For energy consumption, I will not categorize by

sources, such as fossil fuels and renewables, instead I will use the total energy consumption to

focus on the overall trend. I will only use data points from 2006 to 2023, as the types of energy

sources and their ratios are most similar in these years. Moreover, focusing on recent data would

focus the research on the impact of the recent energy consumption trends on the carbon dioxide
1
Hannah Ritchie and Pablo Rosado (2020) - “Energy Mix” Published online at OurWorldinData.org. Retrieved from:
'https://ourworldindata.org/energy-mix' [Online Resource]
2
Andrew, R. (2024, November). Annual carbon dioxide (CO₂) emissions worldwide from 1940 to 2024. Global
Carbon Budget. https://robbieandrew.github.io/GCB2024/

5
emission; making the analysis relevant to current discussion on climate change and energy

policy.

Figure 2: Total global energy consumption by source in TWh against time in years

It is seen from this graph that energy was almost constant from 1800 to 1900, with little

increase by the end of the century. Then, from 1900 to 1950, there was slight increases in energy

consumption. From 1950 onwards, the energy consumption increased significantly but steadily

with some fluctuations along the years. From 2006, “modern biofuels”, “wind”, “solar”, and

other renewables have made visible stronger contributions to energy consumption than in earlier

years; that is why the data I will analyze will be from 2006 to 2023.

6
 Figure 3: Global Carbon dioxide emission in gigaton (Gt) against time in years

Similar to energy consumption graph, the carbon dioxide emissions increased steadily

along the years from 1940 to 2024, with slight fluctuations that may not cause significant effect

on the general trendline.

This table organizes the data from figure 2 and figure 3, matching them according to their

date (year), and organizing them chronologically.

7
 Carbon dioxide emission in Energy consumption in TWh Year

Gigaton (Gt)

37.01 183230 2023

36.50 179819 2022

36.20 176840 2021

34.37 168779 2020

36.37 174458 2019

36.00 172629 2018

35.29 168363 2017

34.73 164719 2016

34.72 163146 2015

34.77 162111 2014

34.65 160629 2013

34.38 158186 2012

33.91 156261 2011

32.81 153125 2010

31.02 146699 2009

31.58 148960 2008

31.60 147434 2007

30.18 143482 2006

Table 1: Carbon dioxide emission and global energy consumption organized by year from 2006
to 2023

8
Exploration
Histograms
Histograms are graphical representations of data that display the frequency of data points

within specified intervals, also known as bins. They are a valuable statistical tool used to

summarize the distribution of a dataset, providing insights into its shape, central tendency,

variability, and potential outliers.

In the context of this investigation, histograms help to visualize the distributions of CO2

emissions and energy consumption over time. By analyzing these distributions, we can identify

patterns such as skewness, uniformity, or clustering of values. This information is crucial for

understanding the nature of the data before applying further statistical methods, such as

correlation or regression analysis. They also provide a visual way to detect irregularities or

anomalies in the data, ensuring its quality and reliability.

Figure 4: histogram of carbon dioxide emissions from 2006 to 2023

9
 The histogram for CO2 emissions shows a concentration of data in the central bin

(32.88–35.58 gigaton), indicating that emissions for most of the years in the dataset were

relatively stable within this range. The distribution appears symmetrical, with fewer occurrences

in the lower (30.18–32.88) and higher (35.58–38.28) ranges. This suggests that global CO2

emissions have hovered around the central range, with slight deviations. If compared with the

number of years, it will be seen that the deviations mostly occur with time, with the first bin

representing the first 5 years (2006-2010). The only exception is in year (2020) with carbon

dioxide emission of (34.37Gt), positioning it in the central bin. This could be attributed to the

economical curtailment due to covid-19 and the quarantine.

Figure 5: histogram of global energy consumption from 2006 to 2023

The histogram for energy consumption exhibits a similar pattern, with the majority of the

data falling within the central range (159,482–175,482 TWh). The leftmost bin (143,482–

159,482 TWh) has slightly more data points, reflecting lower energy usage during the earlier

10
years of the dataset. The rightmost bin (175,482–191,482 TWh) shows a significant drop in

frequency, indicating that only in recent years has energy consumption reached such high levels.

This could be due to industrial growth or increased energy access globally. When compared with

table 1, it is seen that the increase of energy consumption occurs with time, with the first bin

representing the first 7 years, the central bin representing the next 8 years, and the last bin

representing the last 3 years. Although 2020 is not an exception here, its value is far less than the

value of the year previous and subsequent to it, which can be attributed to the same cause for the

decrease of carbon dioxide emissions in 2020.

These histograms help identify trends in the data, providing a visual summary of how

CO2 emissions and energy consumption have varied over time. Their overall alignment between

the ranges, and their increase over time, and even the drop in 2020 supports a positive correlation

between the two variables. The distribution patterns also hint at gradual increases in energy use

and emissions, which will be explored further through regression and correlation analyses.

Scatter plot
Scatter plots are a valuable tool in statistical analysis for examining relationships between

two variables. By plotting individual data points, scatter plots provide a visual representation of

the distribution, patterns, and potential correlations between variables. In this investigation, the

scatter plot illustrates the relationship between Carbon dioxide emissions in gigaton (y-axis) and

energy consumption in terawatt-hours (x-axis) from 2006 to 2023.

11
 39

Carbon dioxide emission (Gt)

37

35

33

31

29

27
140000 145000 150000 155000 160000 165000 170000 175000 180000 185000 190000

Energy consumption (TWh)

Figure 6: Scatter plot, Carbon dioxide emissions against energy consumption

This scatter plot highlights how carbon dioxide emissions and energy consumption vary

together. The upward trend visible in the graph’s trendline suggests a positive association

between the two variables, where higher energy consumption is generally linked to increased

CO2 emissions. Observing the distribution of data points, the trend appears to be predominantly

linear, with CO2 emissions increasing as energy consumption rises. This linear pattern suggests

that as global energy demand grows, CO2 emissions follow a proportional increase, reflecting

the consistent reliance on fossil fuels and other carbon-intensive energy sources. The plot doesn’t

show any significant outliers from the trendline.

The plot does not visibly exhibit characteristics of a quadratic or exponential trend. A

quadratic relationship would show a parabolic shape, either curving upwards or downwards,

while an exponential trend would display a rapid and accelerating increase. In this case, the

relatively steady rate of increase and the uniform spacing of points align more with a linear

relationship.

12
 However, this scatter plot serves as a precursor to deeper statistical analysis. By applying

correlation coefficients and linear regression model, the precise nature of this relationship can be

assessed, validating whether a linear model is indeed the best representation. If not, then other

models will be explored.

Correlation analysis
Correlation analysis is a statistical method used to evaluate the strength and direction of a

relationship between two variables. In the context of this investigation, it helps determine the

degree to which CO2 emissions and energy consumption are related. So, we can understand

whether increases in energy consumption are consistently accompanied by increases in CO2

emissions; quantifying the strength and direction of the relationship. It provides a numerical

measure, the correlation coefficient, which ranges between -1 and 1. A positive coefficient

indicates a direct relationship, where an increase in one variable corresponds to an increase in the

other, while a negative coefficient implies an inverse relationship.

Strength of Prefect Very strong Strong

relationship:

Values of r r =±1 0. 95 ≤ r <1 0. 8 7 ≤ r< 0. 95

−1<r ≤−0.95 −0. 95< r ≤−0. 87

Moderate Weak Very weak No correlation

0. 7 ≤ r <0. 87 0 .5 ≤ r <0. 7 0< r <0.5 r =0

−0. 87< r ≤−0.7 −0. 7<r ≤−0.5 −0.5< r <0

Table 2: strength of correlation and values of r, (Haese, 2019)

13
 For this investigation, correlation analysis is crucial in exploring the research question:

"What is the correlation between carbon dioxide emissions and total energy consumption?" By

applying both Pearson’s correlation (for linear relationships) and Spearman’s rank correlation

(for monotonic relationships), we can comprehensively assess the connection between the two

variables, setting the stage for deeper insights through regression analysis.

Pearson correlation:
I will use Pearson’s correlation coefficient to determine the strength of the relationship

between the 2 variables, using the following formula.

∑ (x i− x̄)( y i − ȳ)
r=
√ ∑ ( x − x̄ ) ∑( y − ȳ)
i
2
i
2

Equation 1: Pearson's correlation coefficient formula (Patil,2023)

Defining key terms:

 r : The correlation coefficient between the two variables (carbon dioxide emissions and

global energy consumption)

 x i: Each year’s global energy consumption

 x̄ : Mean of all the global energy consumption

 y i: Each year’s carbon dioxide emission

 y ̄ : Mean of all Carbon dioxide emission

14
x i and y i are already stated in in table 1. In order to calculate x̄ and ȳ , we will use the following

formula:

∑x
Mean ( x̄)=
n

Equation 2: mean formula

Where n is the number years the data tackles. The equation is substituted by y to get ȳ . The

calculation of both means is shown in table 3:

Mean of the 183230+179819+176840+169779+174458+ …+143482

x̄= x̄=¿162715
18
global
TWh
energy

consumptio

n ( x̄ ¿

Mean of 37.01+36.50+36.20+34.37 +36.37+…+ 30.18

ȳ= ȳ=34.2272> ¿
18
Carbon

dioxide

emission (

ȳ ¿

Table 3: calculating the mean values for global energy consumption and carbon dioxide
emission

To achieve accuracy, all intermediate results will be rounded to 6 significant figures,

except value of r , it will only be 2 S.F. Table 4 will show, step by step, values being substituted

in Pearson’s correlation coefficient formula. Some calculations will only be provided a sample

15
explanation; to avoid repetitiveness and space, the rest of the values will be shown in the

appendix. [1]

Calculation Explaining calculation Result

x i− x̄ Sample for year 2023: (183230−162715) = 20515.0

y i− ȳ Sample for year 2023: (37.01−34.2272 ¿ = 2.78280

(x ¿¿ i− x̄)( y i− ȳ)¿ Sample for year 2023: (20515 ×2.7828) =57089.1

∑ (xi − x̄)( y i− ȳ) 57089.1 ×38873.6 ×27865.5 × … ×77840.2 = 394157

Sum of all values

obtained from
calculations above
2
( x i− x̄ ) Sample for year 2023: (20515)2 =420455000

Squared energy
consumption values
∑ ( x i− x̄ )
2
420455000 × 292547000 ×… ×369908000 =2429970000

Sum of all the squared

energy consumption
values
2
( y i− ȳ) Sample for year 2023: (2.78280)2 =7.74398

Squared carbon dioxide

emission values
∑ ( y i− ȳ )
2
7.74398 ×5.16634 × … ×16.3801 =69.3630

Sum of all squared

carbon dioxide emission
values
∑ ( x i− x̄ ) ( y i− ȳ ) 394157 r =0.96

√∑ ( x − x̄ ) ∑ ( y − ȳ )
i
2
i
2 √(2429970000)×(69.3630)

Calculating r value,
using the above
calculations
Table 4: Calculations of r, by applying Pearson's correlation coefficient formula stepwise

16
 As seen in table 4, r =0.96. This indicates a very strong positive linear relationship

between the two variables, according to table 2. This means that as energy consumption

increases, CO₂ emissions tend to increase as well, almost perfectly. This suggests that as one

variable increases, the other also increases in a nearly predictable manner; the two variables tend

to change together at a consistent rate. This high value implies that the data points closely follow

a straight line, making the relationship highly reliable. It also suggests that knowing the value of

one variable provides a good estimate for the other. However, while the relationship is strong, it’s

important to remember that correlation does not imply causation—it simply indicates a strong

association.

Since the Pearson’s correlation coefficient is very high, and the trendline in the scatter

diagram suggests a clear linear pattern, I will skip spearman’s correlation as it is primarily use

for monotonic relationships and not necessarily linear relationships. There is no evidence, till

now, suggesting a relationship other than linear. So, I will apply Linear regression to quantify the

relationships between the two variables.

Linear regression
Linear regression is a statistical method used to model used to model the relationship

between two variables, by fitting a straight line into the data called line of best fit. In this

investigation, carbon dioxide emissions is the dependent variable while the global energy

consumption is the independent variable. Using linear regression, we can establish a predictive

relationship between them, which will help quantify how changes in energy consumption

impacts carbon dioxide emissions.

17
The regression equation is as follows:

y=mx+b

Equation 3: linear regression equation

Defining key terms:

 y : the dependent variable, carbon dioxide emissions

 m : slope, which is the rate of change

 x : the independent variable, global energy consumption

 b : y -intercept, which represents carbon dioxide emissions when energy consumption is

zero.

To determine line of best fit, we need to calculate slope and y -intercept. To calculate the

slope (m ) use this equation:

∑ ( x i− x̄ ) ( y i − ȳ)
m=
∑ ¿¿

Equation 4: formula to determine slope

Where:

 x i is data points for global energy consumption

 y iis data points for carbon dioxide emissions

 x̄ is mean of global energy consumption

 ȳ is mean of carbon dioxide emissions

18
 ∑ ¿ was calculated in Pearson’s correlation coefficient, and is equal to 2429970000.

∑ (xi − x̄)( y i− ȳ) was also calculated, and is equal to 394157. Therefore, m is equal to:

394157 −4
m= =1.6 ×10
2429970000

Equation 5: substituting values to get the slope

To get y -intercept, we use the following formula:

∑ y−m ∑ x
b=
n

Equation 6: equation of b (y-intercept)

Where:

 n is the number of data points

 ∑ y sum of all y values (carbon dioxide emission)

 ∑ x sum of all x values (global energy consumption)

Value of each variable is shown through the following table:

Variable Value

n 18, as there are 18 data point corresponding to the 18 years calculated (from 2006

to 2023)

∑y ∑ y=37.01+ 36.50+36.20+34.37+ …+30.18=616.0 9

∑x ∑ x=183230+179 819+ 176840+…+143482=293764 0

m −4
1.6 ×10 , as seen from the previous calculations

Table 5: showing value of the y-intercept equation

19
Now, these values will be substituted into the equation:

−4
616.09−(1.6 ×10 × 293764 0)
b= =8.1
18

Equation 7: substituting values into y-intercept equation

According to these calculations, the equation of best-fit line is:

−4
y=1.6 × 10 x +8.1

Equation 8: equation for best-fit line

y -intercept (b ) does not have a real-world meaning or application, since energy

consumption will never be zero. On the other hand, the slope (m ) represents how much carbon

dioxide emissions increase per additional unit of energy consumption. With every 1 TWh, carbon

dioxide emissions increase by 1.6 ×10−4 Gt. Although the change is very small, we must take into

account that a Gigaton is actually a 1 ×109 ton, so the change in 1 TWh increases Carbon dioxide

emissions by 1 .6 ×105 metric tons. Moreover, changes in energy consumption are big, from 2006

to 2023 energy consumption increased by 39748 TWh. This represents an increase by 27.7%.

Energy consumption is rising fast, it is predicted to grow close to 4% annually through 2027.

This would mean an energy consumption of 206198 TWh, and a predicted carbon dioxide

emissions of 42.40 Gt. This increase in carbon dioxide emission will cause catastrophic effects

on the atmosphere and climate change. (calculations are shown in the appendix) [2]

However, statistics show that the peak of carbon dioxide emissions will be reached in

2025 with 39 Gt, and it will steadily decline and drop to 37.4 Gt by 2030. 21.09 Gt isn’t

20
predicted to be reached despite the increase in energy consumption. This is because a linear

relationship doesn’t indicate causation, and linearity may not continue. The model assumes that

the same technology will be used, making it unreliable and inaccurate to a great extent. The

adoption of green and renewable sources of energy is rising rapidly, this would decrease the

carbon dioxide emissions coming from energy consumption, which would make the expected

carbon dioxide emissions an overestimation. It would also count as a warning, that if we didn’t

change the methods of energy production to greener and more renewable ones, the consequences

on earth would be dire. Even though is has its limitations, its strength lies in predicting worst

case scenarios if no changes occurred, acting as a powerful warning.

Conclusion and evaluation

In conclusion, this exploration demonstrated that there is a very strong positive linear

correlation between global energy consumption and carbon dioxide emissions. However, due to

the changing technology, it doesn’t make accurate predictions, as it assumes that the technology

used won’t change to greener methods.

A key strength is the use of multiple statistical methods combined with visual

representations. The histograms used provided insight into the distribution of the variables over

time, showing similar distributions between the two variables. The scatter plot visually

represented the trend, and indicated a positive linear relationship. The Pearson correlation

coefficient confirmed a very strong linear relationship. Applying linear regression quantified the

relationship between the two variables. The dataset used reflects actual global trends, not

21
regional, making the findings applicable to discussions on energy policy and climate change

mitigation.

However, the correlation does not imply causation. Other factors, such as energy policies,

economic activities, energy source and energy efficiency, also influence carbon dioxide

emissions. Different energy sources have different impacts on carbon emissions. Is this

exploration was to be redone, I would explore the effect of fossil fuels as an energy source on

carbon dioxide emissions, focusing only on the most carbon-intensive energy source.

Considering that the aim of the investigation is to understand how energy consumption

patterns affect carbon dioxide emission, the exploration is mostly appropriate. It highlighted the

critical role of energy consumption in influencing carbon dioxide emissions. The strong positive

correlation indicates that energy consumption must be carefully managed to mitigate

environmental impact. As the global demand continues to rise, a shift towards cleaner energy

sources would significantly reduce emissions. This was illustrated in the linear regression

analysis, when the predicted carbon emissions in 2027 was less than the model used predicted,

due to its ignorance to changes in energy sources. This exploration demonstrated the importance

in addressing energy sources while aiming for sustainability. Addressing the energy-emission

link is not just a scientific concern but a public responsibility.

22
Bibliography
 Tiseo, I. (2024, August 21). Global fossil fuel carbon dioxide emissions projections by

sector. Statista. https://www.statista.com/statistics/1385434/fossil-carbon-dioxide-

emissions-projections-by-sector/#:~:text=Global%20fossil%20fuel%20carbon

%20dioxide,to%2037.4%20GtCO%E2%82%82%20by%202030

 International Energy Agency. (2025, February 14). Growth in global electricity demand

is set to accelerate in the coming years as power-hungry sectors expand. IEA.

https://www.iea.org/news/growth-in-global-electricity-demand-is-set-to-accelerate-in-the-

coming-years-as-power-hungry-sectors-expand

 Varsity Tutors. (n.d.). Line of best fit. Retrieved January 15, 2025, from

https://www.varsitytutors.com/hotmath/hotmath_help/topics/line-of-best-fit

 Center for Climate and Energy Solutions. (2021). International emissions. Retrieved

from https://www.c2es.org/content/international-emissions/

 Hannah Ritchie and Pablo Rosado (2020) - “Energy Mix” Published online at

OurWorldinData.org. Retrieved from: 'https://ourworldindata.org/energy-mix' [Online

Resource]

 International Energy Agency. (2023). Energy mix. Retrieved from

https://www.iea.org/world/energy-mix

 Stanford University. (2024). Current energy landscape. Retrieved from

https://understand-energy.stanford.edu/current-energy-landscape

23
  Patil, A. (2023). Pearson's Correlation Coefficient - A Beginners Guide. Analytics

Vidhya. Retrieved December 2, 2024, from

https://www.analyticsvidhya.com/blog/2021/01/beginners-guide-to-pearsons-correlation-

coefficient/

 Haese, M., Humphries, M., Sangwin, C., & Vo, N. (2019). Mathematics: Applications

and Interpretation SL (1st ed.). Haese & Harris Publications.

Appendix
[1]

I organized the data using excel.

24
[2]
 Rate of change (slope) from gigaton to ton is calculated though the equation: m ×109,

substituting the value of the slope (1.6 ×10−4 ×10 9=1.6 ×105 )

 Changes in energy consumption from 2006 to 2023, through the equation:

maximum ( 2023 )−minimum ( 2006 )=183230−143482=39748 TWh

change 39748
 Percentage increase: ¿ ×100 %= ×100 %=27.7 %
initial value (for 2006) 143482

25
  Predicted energy consumption of 2027 with 4% annual increase till then, follows

geometric sequence equation: un =u1 ¿ . n is number of years which is 5 (from 2023 to end

of 2027) and r is the common ratio which is 1.04 (1+0.04), and u1 is value of 2023

(183230). Substituting these values we get:

u5=183230 ¿

 Predicted carbon dioxide emissions for 2027, we will use the equation of linear line

produced in linear regression analysis ( y=1.6 × 10−4 x +8.1) and substitute x with the one

we got from the previous point. y=1.6 × 10−4 ( 214353 )+ 8.1=42.40

26