3 Electricity consumption in United States

In this part, we will explain how social and economic aspects have an impact on the energy consumption. We choose to analyse the most recent values when we want to compare some characteristics to the energy consumption because it is well more representative of the current situation. Furthermore, the electricity consumption is not fluctating that much through the years. In our case, the most recent data are from 2018.

During this exploratory data analysis, we aim to highlight some patterns that explain electricity consumption. We will therefore cover all of these following themes:

  • Energy consumption
  • Population
  • Economy
  • Weather

3.1 Energy consumption

3.1.1 Electricity consumption by states in 2018

Firstly, we are interested in knowing the actual consumption of energy in United States along with its evolving during the last years.

In Figure 3.1, we show the states that consume the most energy in 2018. Without any suprise, the classification looks like the same as the population and GDP classification per state that we will see later. This is obvious that the more populated is a state, the more energy is needed within.

Figure 3.1: Electricity consumption by state in 2018

Figure 3.2: Energy consumption per state [2018]

3.1.2 Electricity consumption per capita in 2018

In order to have a more significant result, we will display the electricity consumption per capita. Indeed, population in states have an important effect on energy consumption.

Figure 3.3: Energy consumption per capita [2018]

In Figure 3.3, we notice that it is not the richest states or the most populated states that consume the most of energy per capita. California and Texas do not consume that much per capita. The new classification is leaded by Wyoming and North Dakota. Wyoming and North Dakota consumed about 30 MWh per capita. It is a huge consumption. To compare, in Switzerland, consumption is arount 7.5 MWh per capita. Other factors should explain this consumption more precisely.

3.1.3 Consumption of electricity depends on its production

It is obvious that the consumption of electricity depends strongly on the avaible energy, namely the electricity production. Underneath, in Figure 3.4, we display this relationship which is really linear. Texas is the stand-out state where the consumption and the production surpass all other states. Observations that can be found above the linear line are consumming more than what they produce and reciprocally for the observations below this line.

Figure 3.4: Comparison between the electricity consumption and the electricity production in 2018

3.1.4 Price of electricity

It is also very interesting to plot the relationship between price of energy and its consumption as we can see in Figure 3.5. But we can see that the energy consumption does not depend on the price because energy is a necessity good. Nowadays, we can no longer deprive ourselves of it.

Figure 3.5: Price of the energy in comparison to the consumption in 2018

3.1.5 Evolution of the overall electricity consumption

By plotting the evolution of the electricity consumption in the United States, we can see that the trend is growing through the years but that the economic crisis in 2008 has reduced the electricity consumption. However, the trend has gone up since the crisis.

Figure 3.6: Evolution of the electricity consumption in US

3.2 Population

3.2.1 Population per state in 2018

In Table 3.1, we display a list and in Figure 3.7 a map of states with their population (2018). This is an important variable in order to explain the electricity consumption. It is evident that states having more population will consume more energy. We want to know the most populated states.

Table 3.1: Population by state [2018]
State population
California 39’461’588
Texas 28’628’666
Florida 21’244’317
New York 19’530’351
Pennsylvania 12’800’922
Illinois 12’723’071
Ohio 11’676’341
Georgia 10’511’131
North Carolina 10’381’615
Michigan 9’984’072
New Jersey 8’886’025
Virginia 8’501’286
Washington 7’523’869
Arizona 7’158’024
Massachusetts 6’882’635
Tennessee 6’771’631
Indiana 6’695’497
Missouri 6’121’623
Maryland 6’035’802
Wisconsin 5’807’406
Colorado 5’691’287
Minnesota 5’606’249
South Carolina 5’084’156
Alabama 4’887’681
Louisiana 4’659’690
Kentucky 4’461’153
Oregon 4’181’886
Oklahoma 3’940’235
Connecticut 3’571’520
Utah 3’153’550
Iowa 3’148’618
Nevada 3’027’341
Arkansas 3’009’733
Mississippi 2’981’020
Kansas 2’911’359
New Mexico 2’092’741
Nebraska 1’925’614
West Virginia 1’804’291
Idaho 1’750’536
Hawaii 1’420’593
New Hampshire 1’353’465
Maine 1’339’057
Montana 1’060’665
Rhode Island 1’058’287
Delaware 965’479
South Dakota 878’698
North Dakota 758’080
Alaska 735’139
District of Columbia 701’547
Vermont 624’358
Wyoming 577’601

Figure 3.7: Map of population per state in 2018

Partly because California and Texas are the most populated states, they consume more energy than other ones.

3.2.2 Population in 2018 in comparison to the annual consumption of electricity

In Figure 3.8, we want to confirm the relationship between population and electricity consumption. Indeed, there is a trend between these two variables.

Figure 3.8: Comparison of the popoulation size and the annual electricity consumption in 2018

3.2.3 Density compared to the annual consumption per state in 2018

Figure 3.9 shows that states with a density between 0 and 100 inhabitants per square kilometer follow a linear trend. After this threshold, density can no longer explain the consumption of electricity. Also, we decided to remove the District of Columbia because it is an outlier.

Figure 3.9: Comparison between the density of the sates and the annual electricity consumption in 2018

3.2.4 Number of houses compared to the consumption of electricity.

The linear trend in Figure 3.10 is strong. The more households there are, the more electricity is consumed. Few states stand out from the rest. This fact was quite obvious but needed to be verifed.

Figure 3.10: Comparison between the number of houses and the electricity consumption in 2018

3.2.5 Number of capita per house in comparison to the electricity consumption.

The relationship between number of capita per house and electricity consumption is not really significant in Figure 3.11. The consumption of electricity does not depend on the number of people per house.

Figure 3.11: Comparison between the number of capita per house and the electricity consumption

3.3 Economy

In this part, we will compare economic characteristics and energy consumption.

3.3.1 GDP in states

The US states producing the most wealth in 2018 are listed in the Table 3.2.

Table 3.2: GDP by state in dollars [2018]
State total_GDP_trillion GDP_per_capita
California 3,00 75’966
Texas 1,80 62’962
New York 1,67 85’450
Florida 1,04 48’918
Illinois 0,87 68’011
Pennsylvania 0,78 61’181
Ohio 0,68 57’887
New Jersey 0,62 69’998
Georgia 0,59 56’336
Massachusetts 0,57 82’743
Washington 0,57 75’205
North Carolina 0,56 54’297
Virginia 0,53 62’684
Michigan 0,53 52’794
Maryland 0,41 68’356
Colorado 0,37 65’319
Minnesota 0,37 65’793
Indiana 0,37 54’783
Tennessee 0,36 53’769
Arizona 0,35 48’658
Wisconsin 0,34 57’908
Missouri 0,32 52’097
Connecticut 0,28 77’202
Louisiana 0,26 55’216
Oregon 0,24 57’338
South Carolina 0,23 46’012
Alabama 0,22 45’366
Kentucky 0,21 46’644
Oklahoma 0,20 51’407
Iowa 0,19 60’249
Utah 0,18 56’488
Nevada 0,17 55’927
Kansas 0,17 57’814
District of Columbia 0,14 200’501
Arkansas 0,13 42’668
Nebraska 0,12 64’384
Mississippi 0,11 38’522
New Mexico 0,10 47’926
Hawaii 0,09 66’027
New Hampshire 0,08 62’406
West Virginia 0,08 42’919
Idaho 0,08 44’016
Delaware 0,07 76’109
Maine 0,06 48’434
Rhode Island 0,06 57’251
North Dakota 0,06 73’979
Alaska 0,05 74’454
South Dakota 0,05 59’195
Montana 0,05 47’448
Wyoming 0,04 67’726
Vermont 0,03 53’265

We can see that the total population by state is strongly correlated with the creation of wealth. Indeed, the most populated states, namely California, Texas, Florida and New York, are also the ones that produce the most wealth (Figure 3.12).

Figure 3.12: GDP per state in 2018

Four states exceed one trillion. California even reaches 3 trillion. By way of comparison, Switzerland would be 7th in this ranking with a GDP of 705 billion CHF. Today, creating wealth means consuming energy. This variable could well explain the consumption of states as well as the price of electricity.

3.3.2 GDP versus energy consumption

In Figure 3.13, we display the comparison between GDP and energy consumption per state in 2018.

Figure 3.13: Comparison between GDP and electricity consumption in 2018

In point of fact, energy consumption in well related to GDP.

3.3.3 Comparison between the variation of GDP and the variation of electricity production through the years.

In Figure 3.14, we display the variation of GDP and the variation of electricity consumption from years 2000.

Figure 3.14: Comparison between the variation in consumption and the variation in GDP

We can see that US economic activity influences energy consumption. The similarity between the two curves is clear during the last 10 years. The 2008 crisis led to a drop in GDP of almost 4% and a drop in energy consumption of around 6%. However, we have to pay attention to the fact that correlation does not mean causation. Here, the decrease of GDP probably lowers the production of electricity and thereby the electricity consumption.

3.3.4 GDP per capita

In terms of GDP per capita, we see great inequality. While it reaches 85,450 dollars in New York state, it is only 38,550 dollars in Mississippi state. This trend could also be seen in energy consumption per capita.

There is no trend between GDP per capita and energy consumption. The majority of values are stagnating between 0 and 200, irrespective of their GDP per capita. However, we notice the presence of an outlier which is the district of Columbia. This is because this small district has a lot of high value-added jobs in the administration and a small population. Again, for this part, we remove this district because it distorts our graphic.

The largest economies are not the richest per capita. Indeed, California and Texas have the biggest GDP but also have many people. Despite the fact that most populated states have the bigget GDP, they are not necessary the richest ones. This is the reason why the GDP per capita does not follow the trend of the GDP and then, that there is no relationship between GDP per capita and energy consumption.

As a conclusion, this is surely a problem of collinerity between GDP and population. Thus, GDP or population have maybe no real relationship with electricity consumption. We will try to verify this assumption in the modelling part.

Figure 3.15: Comparison between GDP per capita and electricity consumption in 2018

3.3.5 Personal income in 2018

In Figure 3.16, the personal income does not explain energy consumption. Indeed, electricity is a basic commodity and the increase in revenue will not significantly increase electricity consumption.

Figure 3.16: Comparison between the personal income and the electricity consumption in 2018

3.3.6 Buildings

In Figure 3.17, we are showing the repartition between industrial, commercial and residential buildings (customers) in terms of electricity consumption in 2018.

Figure 3.17: Proportion in residential, industrial and commercial sector in each state in 2018

In order to have a better view of the proportion of industrial customers, we will plot it separately in the next graph (Figure 3.18).

Figure 3.18: Proportion of residential customers by states in 2018

Nebraska, Idaho, Whyoming, Arkansas and North Dakota are the most industrial states. Then, it can explain the important energy consumption per capita in those states and particularly in Wyoming and North Dakota. Thus, this variable will be quite important to predict the consumption of energy.

In the following graphs, (Figure 3.19, Figure 3.20, Figure 3.21), we display the number of customers in each category.

Figure 3.19: Number of residential customers in each state in 2018

Figure 3.20: Number of industrial customers in each state in 2018

Figure 3.21: Number of commercial customers in each state in 2018

Thanks to those different previous graphics, we see that California, Texas, Florida and New-York have the more customers in nearly each sector. Thus, they have lots of people within but also a strong industrial and commercial activity. Consequently, they produce more GDP and also need more energy than other states.

Then, we compare the customers of each sector and the consumption of electricity.

Figure 3.22: Comparison between the number of residential customers and the electricity consumption

Figure 3.23: Comparison between the number of industrial customers and the electricity consumption

Figure 3.24: Comparison between the number of commercial customers and the electricity consumption

As we can see in the previous plots (Figure 3.22, Figure 3.23, Figure 3.24), the type of building is essential to know the energy consumption and especially for residential and commercial buildings. Thus, states having more residential or/and commercial customers tend to consume more than others.

3.4 Weather

Temperature generally plays a role in the consumption and generation of electricity. The United States is a country with a very varied climate.

Do we notice differences in production and consumption according to the state and temperatures ? (Warmer temperatures so more air conditioner/colder temperatures so more heater for example ?)

Finally, sunlight also impacts the electricity sector. For example, it plays a role on the production of solar panels, but also on energy consumption (lighting, heating, etc.). In the following, we will verify all these assumptions.

Figure 3.25: Summer temperatures in 2018

Figure 3.26: Comparison between summer temperatures and electricity consumption in 2018

Figure 3.27: Winter temperatures in 2018

Figure 3.28: Comparison between winter temperatures and electricity consumption in 2018

Firstly, in those two different maps (Figure 3.25 and 3.27), the summer one and the winter one, we can see that the country is clearly divided into two parts: the south part with higher temperatures in winter and summer and the north part with lower temperatures.

However, the relationship between temperature and electricity consumption is quite weak but still more pronounced during the winter season (Figure 3.28) than the summer one (Figure 3.26). Meanwhile, we do not detect any link with the most consumer states. But, a trend could be stronger for the production or the pricing of electricity.

In Figure 3.29, we display the map of the number or hours of sun in US. Again, we look for a pattern between hours of sun and consumption of electricity.