California forest fires acres battled by agency

forest fires

An exercise in effective use of color and axis labels

Published

July 19, 2025

My interpretation of forest fire data courtesy of CalFire. This page is currently undergoing updates.

Libraries

For two reasons (the fact that our data lives in a PDF file plus a few other sources, and the fact that I’m using a pre-defined color palette) we need three packages in addition to the tidyverse in order to produce the plot.

palettetown: a package full of color palettes derived from Pokemon and inspired by pokepalettes; each Pokemon is represented in the package, but my favorite is Charizard and happens to go along with the theme of the data.
pdftools: this data is stored in a pdf file, so this is an opportunity to show how to use R to replicate the process of extracting the data from the file and into a data frame from which to be plotted.
janitor: a very helpful package for restructuring dirty data frames, such as those extracted from pdf files.

Code

library(tidyverse)
library(palettetown)
library(pdftools)
library(janitor)

Data Processing

To begin the process of extracting the data from the pdf, let’s instantiate some objects.

Code

pdf = "pdfs/fires-acres-all-agencies-thru-2018.pdf"
tib = pdf_data(pdf)[[1]] 

tib

# A tibble: 355 × 6
   width height     x     y space text         
   <int>  <int> <int> <int> <lgl> <chr>        
 1    88     22   153    18 TRUE  California   
 2    80     22   246    18 TRUE  Wildfires    
 3    32     22   331    18 TRUE  and          
 4    48     22   368    18 TRUE  Acres        
 5    25     22   422    18 TRUE  for          
 6    21     22   453    18 TRUE  all          
 7   110     22   479    18 FALSE Jurisdictions
 8    18      9   145    43 TRUE  CAL          
 9    21      9   167    43 FALSE FIRE         
10    26      9   116    57 FALSE FIRES        
# ℹ 345 more rows

Code

partial = tib %>% 
  .[23:278, c(4,6)] %>% 
  group_by(y) %>% 
  nest() %>% 
  pull(data) %>% 
  bind_cols() %>% 
  row_to_names(1) %>% 
  add_column(source = c("fires_cal",
                        "acres_cal",
                        "fires_fed",
                        "acres_fed",
                        "fires_local",
                        "acres_local",
                        "fires_total"),
             .before = "1987") %>% 
  pivot_longer(cols = -1) %>% 
  pivot_wider(names_from = "source",
              values_from = "value") %>% 
  dplyr::rename(year = name)

more = tib %>% 
  .[323:354, 6] %>% 
  select(acres_total = text)

df = partial %>% 
  bind_cols(more) %>% 
  mutate(across(2:9, ~ as.numeric(eeptools::decomma(.))),
         year = as.numeric(partial$year)) 

rest = tibble(
  year = c(2019, 2020, 2021, 2022, 2023),
  fires_cal = c(3086, # 2019 value
                3501, # 2020 value 
                3054, # ...
                3333, 
                3236),
  acres_cal = c(129914, # 2019 value
                1458881, # 2020 value 
                279703, # ... 
                70933, 
                24971),
  fires_fed = c(997 + 156 + 34 + 80 + 15 + 2, # 2019 value
                1421 + 217 + 79 + 97 + 13 + 5, # 2020 value 
                1267 + 183 + 115 + 83 + 20 + 1, # ...
                934 + 79 + 50 + 36 + 13 + 19,
                1022 + 82 + 84 + 65 + 8 + 7),
  acres_fed = c(1112399 + 8539 + 111+334 + 2754 + 33, # 2019 value
                2520946 + 142201 + 76796 + 22210 + 45 + 11476, # 2020 value 
                2029239 + 30145 + 109420 + 98793 + 552 + 1000, # ...
                10932 + 234624 + 3240 + 6864 + 156 + 193 + 128,
                187255 + 1557 + 90251 + 201 + 8),
  fires_local = c(2370 + 408,  # 2019 value
                  2849 + 466, # 2020 value 
                  2420 + 253, # ... 
                  2642 + 371, 
                  2508 + 374),
  acres_local = c(7220 + 15981,  # 2019 value
                  17062 + 54762, # 2020 value 
                  13828 + 6706, # ... 
                  4288 + 10932, 
                  4936 + 18976),
  fires_total = c(7148, 8648, 7396, 7490, NA),
  acres_total = c(277285, 4304379, 2569386, 362455, NA)
)

fires <- 
  df %>% 
  rbind(rest) %>% 
  pivot_longer(cols = fires_cal:acres_total, names_to = "type") %>% 
  filter(str_starts(type, "acres") & !str_ends(type, "total")) %>% 
  mutate(
    type = case_when(
      type == "acres_local" ~ "Local",
      type == "acres_cal" ~ "State",
      type == "acres_fed" ~ "Federal"
    ) %>% 
      as.factor(.) %>% 
      recode_factor(
        ., 
        "Local" = "Local",
        "State" = "State",
        "Federa" = "Federal"))

Color scheme

Pulling our color scheme from the palettetown package that produces hex codes corresponding with first-generation Pokemon. We’ll select Charizard.

Code

char <- palettetown::ichooseyou("charizard")
extrafont::loadfonts()

Plotting

Code

p <- 
  fires %>% 
  ggplot(aes(x = year, y = value/1000, fill = type)) +
  geom_col(color = "white",
           width = 1) +
  labs(title = "Califoria forest fires acres battled by agency",
       subtitle = "Source: CalFire",
       x = "",
       y = "",
       fill = "",
       caption = "solplots.com") +
  scale_fill_manual(values = c("Federal" = char[11],
                               "State" = char[8],
                               "Local" = char[2])) +
  scale_x_continuous(expand = c(0, 0),
                     breaks = seq(1987, 2023, 4)) +
  scale_y_continuous(expand = c(0, 0),
                     limits = c(0, 4700)) +
  annotate("text", 
           x = rep(1985, 5),
           y = seq(150, 4150, 1000),
           label = c(0:3, "4 million acres"),
           color = "white",
           family = "Garamond",
           hjust = 0,
           size = 5) +
  theme(panel.grid.minor.x = element_blank(),
        panel.grid.major.x = element_blank(),
        panel.grid.minor.y = element_blank(),
        panel.grid = element_line(color ="white"),
        panel.border = element_blank(),
        plot.title = element_text(size = rel(1.3),
                                  face = "bold"),
        plot.subtitle = element_text(size = rel(1.2)), 
        text = element_text(color = "white",
                            family = "Georgia"),
        axis.text.y = element_blank(),
        axis.text.x = element_text(color = "white",
                                   size = rel(1.3)),
        panel.background = element_rect(fill = char[4]),
        plot.background = element_rect(fill = char[4]),
        legend.background = element_rect(fill = char[4],
                                         color = char[4]),
        legend.title = element_blank(),
        legend.text = element_text(size = rel(1.1)),
        legend.position = "inside",
        legend.position.inside = c(0.26, 0.965),
        legend.direction = "horizontal",
        axis.ticks = element_blank(),
        plot.caption = element_text(face = "bold",
                                    size = 10),
        plot.margin = unit(c(5,5,3,0), "mm"))

p

--- title: "California forest fires acres battled by agency" toc: true date: "2025-07-19" description: "An exercise in effective use of color and axis labels" categories: [forest fires] image: "images/fighting_acres_2025.png" format: html: code-fold: true code-tools: true execute: warning: false message: false --- My interpretation of forest fire data courtesy of [CalFire](https://www.fire.ca.gov/our-impact/statistics). This page is currently undergoing updates. # Libraries For two reasons (the fact that our data lives in a PDF file plus a few other sources, and the fact that I'm using a pre-defined color palette) we need three packages in addition to the tidyverse in order to produce the plot. - [palettetown](https://github.com/timcdlucas/palettetown?tab=readme-ov-file): a package full of color palettes derived from Pokemon and inspired by [pokepalettes](https://pokepalettes.com/#charizard); each Pokemon is represented in the package, but my favorite is Charizard and happens to go along with the theme of the data. - [pdftools](https://github.com/ropensci/pdftools): this data is stored in a pdf file, so this is an opportunity to show how to use R to replicate the process of extracting the data from the file and into a data frame from which to be plotted. - [janitor](https://github.com/sfirke/janitor): a very helpful package for restructuring dirty data frames, such as those extracted from pdf files. ```{r} library(tidyverse) library(palettetown) library(pdftools) library(janitor) ``` # Data Processing To begin the process of extracting the data from the pdf, let's instantiate some objects. ```{r} pdf = "pdfs/fires-acres-all-agencies-thru-2018.pdf" tib = pdf_data(pdf)[[1]] tib ``` ```{r} partial = tib %>% .[23:278, c(4,6)] %>% group_by(y) %>% nest() %>% pull(data) %>% bind_cols() %>% row_to_names(1) %>% add_column(source = c("fires_cal", "acres_cal", "fires_fed", "acres_fed", "fires_local", "acres_local", "fires_total"), .before = "1987") %>% pivot_longer(cols = -1) %>% pivot_wider(names_from = "source", values_from = "value") %>% dplyr::rename(year = name) more = tib %>% .[323:354, 6] %>% select(acres_total = text) df = partial %>% bind_cols(more) %>% mutate(across(2:9, ~ as.numeric(eeptools::decomma(.))), year = as.numeric(partial$year)) rest = tibble( year = c(2019, 2020, 2021, 2022, 2023), fires_cal = c(3086, # 2019 value 3501, # 2020 value 3054, # ... 3333, 3236), acres_cal = c(129914, # 2019 value 1458881, # 2020 value 279703, # ... 70933, 24971), fires_fed = c(997 + 156 + 34 + 80 + 15 + 2, # 2019 value 1421 + 217 + 79 + 97 + 13 + 5, # 2020 value 1267 + 183 + 115 + 83 + 20 + 1, # ... 934 + 79 + 50 + 36 + 13 + 19, 1022 + 82 + 84 + 65 + 8 + 7), acres_fed = c(1112399 + 8539 + 111+334 + 2754 + 33, # 2019 value 2520946 + 142201 + 76796 + 22210 + 45 + 11476, # 2020 value 2029239 + 30145 + 109420 + 98793 + 552 + 1000, # ... 10932 + 234624 + 3240 + 6864 + 156 + 193 + 128, 187255 + 1557 + 90251 + 201 + 8), fires_local = c(2370 + 408, # 2019 value 2849 + 466, # 2020 value 2420 + 253, # ... 2642 + 371, 2508 + 374), acres_local = c(7220 + 15981, # 2019 value 17062 + 54762, # 2020 value 13828 + 6706, # ... 4288 + 10932, 4936 + 18976), fires_total = c(7148, 8648, 7396, 7490, NA), acres_total = c(277285, 4304379, 2569386, 362455, NA) ) fires <- df %>% rbind(rest) %>% pivot_longer(cols = fires_cal:acres_total, names_to = "type") %>% filter(str_starts(type, "acres") & !str_ends(type, "total")) %>% mutate( type = case_when( type == "acres_local" ~ "Local", type == "acres_cal" ~ "State", type == "acres_fed" ~ "Federal" ) %>% as.factor(.) %>% recode_factor( ., "Local" = "Local", "State" = "State", "Federa" = "Federal")) ``` # Color scheme Pulling our color scheme from the `palettetown` package that produces hex codes corresponding with first-generation Pokemon. We'll select Charizard. ```{r} char <- palettetown::ichooseyou("charizard") extrafont::loadfonts() ``` # Plotting ```{r, fig.width = 5.1, fig.height = 5.1} p <- fires %>% ggplot(aes(x = year, y = value/1000, fill = type)) + geom_col(color = "white", width = 1) + labs(title = "Califoria forest fires acres battled by agency", subtitle = "Source: CalFire", x = "", y = "", fill = "", caption = "solplots.com") + scale_fill_manual(values = c("Federal" = char[11], "State" = char[8], "Local" = char[2])) + scale_x_continuous(expand = c(0, 0), breaks = seq(1987, 2023, 4)) + scale_y_continuous(expand = c(0, 0), limits = c(0, 4700)) + annotate("text", x = rep(1985, 5), y = seq(150, 4150, 1000), label = c(0:3, "4 million acres"), color = "white", family = "Garamond", hjust = 0, size = 5) + theme(panel.grid.minor.x = element_blank(), panel.grid.major.x = element_blank(), panel.grid.minor.y = element_blank(), panel.grid = element_line(color ="white"), panel.border = element_blank(), plot.title = element_text(size = rel(1.3), face = "bold"), plot.subtitle = element_text(size = rel(1.2)), text = element_text(color = "white", family = "Georgia"), axis.text.y = element_blank(), axis.text.x = element_text(color = "white", size = rel(1.3)), panel.background = element_rect(fill = char[4]), plot.background = element_rect(fill = char[4]), legend.background = element_rect(fill = char[4], color = char[4]), legend.title = element_blank(), legend.text = element_text(size = rel(1.1)), legend.position = "inside", legend.position.inside = c(0.26, 0.965), legend.direction = "horizontal", axis.ticks = element_blank(), plot.caption = element_text(face = "bold", size = 10), plot.margin = unit(c(5,5,3,0), "mm")) p ```