Exploring Energy Capacity and Prices

1 Loading the data

This post revisits an older Tidy Tuesday dataset, it is concerned with capacity and costs for solar and wind energy.

pacman::p_load(tidyverse)
capacity <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2022/2022-05-03/capacity.csv')
wind <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2022/2022-05-03/wind.csv')
solar <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2022/2022-05-03/solar.csv')
average_cost <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2022/2022-05-03/average_cost.csv')

2 Cleaning the data

Let’s take a look at the average cost data:

average_cost %>% 
  head(5)

# A tibble: 5 × 4
   year gas_mwh solar_mwh wind_mwh
  <dbl>   <dbl>     <dbl>    <dbl>
1  2009    57.6     168.      74.3
2  2010    56.8     140.      65.5
3  2011    46.0     111.      47.8
4  2012    44.5      84.1     40.1
5  2013    43.2      68.9     28.7

In order to make visualisation easier we are going to convert the data into long format, using the pivot_longer function:

average_cost %>% pivot_longer(cols = gas_mwh:wind_mwh,names_to = "type",values_to = "cost") %>%  mutate(flag = if_else(type == 'solar_mwh', 'top', false = 'bottom')) -> average_cost #Assign to a new variable

average_cost %>% 
  select(-flag)

# A tibble: 39 × 3
    year type       cost
   <dbl> <chr>     <dbl>
 1  2009 gas_mwh    57.6
 2  2009 solar_mwh 168. 
 3  2009 wind_mwh   74.3
 4  2010 gas_mwh    56.8
 5  2010 solar_mwh 140. 
 6  2010 wind_mwh   65.5
 7  2011 gas_mwh    46.0
 8  2011 solar_mwh 111. 
 9  2011 wind_mwh   47.8
10  2012 gas_mwh    44.5
# … with 29 more rows

The wrangled data is now far more suitable for visualisation.

3 Visualise the data

Let’s visualise the data looking at the costs over time for each of the energy sources.

average_cost %>% ggplot(aes(year,cost,color=type))+
  geom_line(aes(linetype=flag), size=1.5,alpha =0.8)+
  scale_x_continuous(breaks = seq(2008, 2022, by = 2))+
  scale_y_continuous(labels = scales::dollar_format(), breaks = seq(0,175,by=25))+
  tidyquant::theme_tq()+
  expand_limits(y=0)+
  guides(linetype="none")+
  scale_linetype_manual(values = c("dashed","solid"))+
  labs(
    x="",
    y="Cost",
    title = "Energy costs over the years",
    subtitle = "Solar has experienced the most drastic decrease in costs",
    caption = "Data from Tidy Tuesday: 04-05-2022")+
    tidyquant::theme_tq()+
    tidyquant::scale_color_tq()+
    theme(legend.title = element_blank()) -> plot1

plot1

We can see that in solar has experienced the most drastic decrease in costs for the available date ranges in the data. Let’s get a sense of the spread of the prices for solar:

solar %>% 
  ggplot(
    aes(solar_mwh)
  )+
  geom_histogram(color = 'white',
                 fill ='midnightblue')+
  geom_rug()+
  labs(
    title = "Distribution of price for solar",
    x = "",
    y= "Count"
  ) +
  scale_x_continuous(labels = scales::dollar_format(suffix = "/MWh")) +
  # scale_y_continuous(expand = c(0,0), limits = c(0,50))+
  tidyquant::theme_tq() -> plot2

plot2

The prices for solar energy vary from 12/MWh to just above 300/MWh, the majority between 12 and 100. Let’s also try to understand the relation ship between $/MWh and solar capacity.

solar %>% 
  ggplot(aes(solar_capacity,solar_mwh, color=solar_capacity))+
  geom_jitter()+
  geom_smooth(color='midnightblue')+
  tidyquant::theme_tq()+
  theme(legend.position = "none")+
  scale_color_continuous(name = 'Wind Capacity') +
  scale_y_continuous(labels = scales::dollar_format(suffix = "/MWh"))+
  labs(
    x="Solar Capacity",
    y="Solar Projected Price",
    title = "Relationship between solar\nprojected price and capacity") -> plot3

plot3

The relationship is not too strong, but at an overall level we can say that the price decreases as solar capacity increases, up to a certain point where it seems to stabilise.

4 Bring it all together

Let’s bring all of these visualisations together using the patchwork package:

library(patchwork)

((plot2/plot3|plot1))