Rachel_4_19_Presentation.R

### Materials for Rachel Weber's 4/19/24 Slide Deck ###
### MVH 4-12-24 ###

# Load packages and data

library(tidyverse)

options(scipen = 999)

# Trends in Incentivized FMV as a percent of the base over time

## Data prep

muni_MC <- read_csv("./Output/ptaxsim_muni_class_summaries_2006-2022.csv") %>%
  select(year, clean_name, class, av)

class_dict <- read_csv("./Necessary_Files/class_dict_expanded.csv") %>%
  select(class = class_code, class_1dig, assess_ratio, incent_prop, Alea_cat, major_class_code)

muni_MC <- muni_MC %>% 
  left_join(class_dict, by = c("class")) %>% 
  filter(class !=0) # drop exempt property types with 0 taxable value

#class_8_munis <- read_csv("./Necessary_Files/datarequests_Class8Munis.csv")
class_8_munis <- read_csv("./Output/datarequests_Class8Munis.csv")

# changed from as.list to as.character
class_8_munis <- as.character(class_8_munis$clean_name)



# class 8 munis - at the year-class level
class_8_df <- # left_join(muni_MC, class_dict, by = "class") %>%
  muni_MC %>%
  filter(clean_name %in% class_8_munis) %>%
  filter(av != 0) %>%
  mutate(FMV = av/assess_ratio) %>%
  group_by(year) %>%   
  mutate(year_tb_tot = sum(FMV)) %>%           # tax base for all class 8 munis together, per year
  ungroup() %>%
  filter(Alea_cat %in% c("Industrial", "Commercial")) %>%      ## drops all non-industrial and non-commercial classes to calculate the rest of the totals
  group_by(year) %>%
  mutate(year_ind_comm_FMV = sum(FMV)) %>%    #  total commercial and industrial FMV for all class 8 munis together, per year
  ungroup() %>%
  group_by(year, clean_name) %>%
  mutate(muni_year_ind_comm_FMV = sum(FMV)) %>%   # calculates total FMV in each munis  each year
  ungroup() %>%
  group_by(year, Alea_cat) %>%
  mutate(cat_year_FMV = sum(FMV)) %>%    # calculates FMV within each commercial vs industrial category for each year
  ungroup() %>%
  
  ## Do you want class_1dig or major_class_code? Does it matter? 
  ## probably not if I already had joined in the commecial and industrial codes to the class level data...
  group_by(year, clean_name, class_1dig) %>%   # total fmv in class 5, 6, 7, and 8 per muni
  mutate(year_muni_class_FMV = sum(FMV))

## Added this for cook level totals:
class_8_df_outofCook <- 
  muni_MC %>%
 # filter(clean_name %in% class_8_munis) %>% ## keep all munis, use for cook county totals.
  filter(av != 0) %>%
  mutate(FMV = av/assess_ratio) %>%
  group_by(year) %>%   
  mutate(year_tb_tot = sum(FMV)) %>%           # tax base for all class 8 munis together, per year
  ungroup() %>%
  filter(Alea_cat %in% c("Industrial", "Commercial")) %>%      ## drops all non-industrial and non-commercial classes to calculate the rest of the totals
  group_by(year) %>%
  mutate(year_ind_comm_FMV = sum(FMV)) %>%    #  total commercial and industrial FMV for all class 8 munis together, per year
  ungroup() %>%
  group_by(year, clean_name) %>%
  mutate(muni_year_ind_comm_FMV = sum(FMV)) %>%   # calculates total FMV in each munis  each year
  ungroup() %>%
  group_by(year, Alea_cat) %>%
  mutate(cat_year_FMV = sum(FMV)) %>%    # calculates FMV within each commercial vs industrial category for each year
  ungroup() %>%
  
  ## Do you want class_1dig or major_class_code? Does it matter? 
  ## probably not if I already had joined in the commecial and industrial codes to the class level data...
  group_by(year, clean_name, class_1dig) %>%   # total fmv in class 5, 6, 7, and 8 per muni
  mutate(year_muni_class_FMV = sum(FMV))

 class(class_8_df$class_1dig)   ## moved this lower because it didn't exist yet when you ran it 

# Class 8 Munis

## Incentive types over time (line graph and line-area graph)

# ggplot() +
#   geom_line(data = class_8_df %>%
#              summarize(ind_comm_perc = year_ind_comm_FMV/year_tb_tot), aes(x = year, y = ind_comm_perc)) +#, color = "Ind. & Comm. FMV") +
#   geom_line(data = class_8_df %>%
#               filter(incent_prop == "Incentive") %>%
#               group_by(year) %>%
#               summarise(incent_perc = sum(FMV)/year_tb_tot),
#             aes(x = year, y = incent_perc)#,
#             #color = "Incent. Class FMV"
#             ) +
#   geom_line(data = class_8_df %>%
#               filter(class_1dig == 8)) +        # was missing a + sign here
#   theme_classic() +
#   scale_x_continuous(breaks = seq(2006, 2022, by = 3)) +
#   scale_y_continuous(labels = scales::percent_format(), limits = c(0, 0.45), breaks = seq(0, 0.5, by = 0.05))

 
## Class 8 Townships Graph -------------------
 ## Alea Version: 
ggplot() +
  geom_line(data = class_8_df %>%
              group_by(year) %>%
              summarize(ind_comm_perc = mean(year_ind_comm_FMV/year_tb_tot)), 
            aes(x = year, y = ind_comm_perc, color =  "Commercial+Industrial"), lwd = 1) +  
  
  # industrial fmv
  geom_line(data = class_8_df %>%               
              filter(Alea_cat == "Industrial") %>%
              group_by(year) %>%
              # needed na.rm=TRUE, otherwise it didn't work. perc_industrial was not being calculated without it.
              summarize(perc_industrial =  sum(FMV/year_tb_tot, na.rm=TRUE)),   ## added this part
            aes(x = year, y = perc_industrial, color = "Industrial"), lwd = 1) +
  
  # commercial fmv
  geom_line(data = class_8_df %>%              
              filter(Alea_cat == "Commercial") %>%
              group_by(year) %>%
              summarize(perc_commercial =  sum(FMV/year_tb_tot, na.rm=TRUE)),   ## added this part
            aes(x = year, y = perc_commercial, color = "Commercial"), lwd = 1) +
  geom_line(data = class_8_df %>%
              filter(incent_prop == "Incentive") %>%
              group_by(year) %>%
              summarise(incent_perc = sum(FMV)/year_tb_tot),
            aes(x = year, y = incent_perc, color = "Incentive Classes"), lwd = 1 ) +
  geom_line(data = class_8_df %>%               # threw error here, missing x and y in aes()
              filter(class_1dig == 8) %>% 
              group_by(year) %>%
              summarize(perc_8 =  sum(FMV/year_tb_tot)),   ## added this part
            aes(x = year, y = perc_8, color = "Class 8"), lwd = 1) +        # was missing a + sign here
  theme_classic() +
  scale_x_continuous(name = "", breaks = seq(2006, 2022, by = 3), limits = c(2006, 2022), expand = c(0,0)) +
  scale_y_continuous(name = "Percent of FMV", labels = scales::percent_format(), limits = c(0, 0.20),
                     breaks = seq(0, 0.5, by = 0.05), expand = c(0,0))  +
 scale_color_manual(name = "", values = c("Commercial+Industrial" = "black", "Industrial" = "gray70", "Commercial" = "gray50",  "Incentive Classes" = "orange",  "Class 8" = "red" )) +
  theme(legend.position = "bottom") + 
  labs(title = "Property in the Class 8 Townships") + 
  guides(color = guide_legend(nrow=2, byrow = TRUE))


## Percentage out of Cook County ---------------------
 
## Alea Version for Cook Level: 
ggplot() +
  # Commercial + Industrial FMV in cook
  geom_line(data = class_8_df_outofCook %>%
              group_by(year) %>%  #didn't group by year before?
              summarize(ind_comm_perc = mean(year_ind_comm_FMV/year_tb_tot)), 
            aes(x = year, y = ind_comm_perc, color =  "Commercial+Industrial"), lwd = 1) +  
  
  # incentive class properties in cook
  geom_line(data = class_8_df_outofCook %>%
              filter(incent_prop == "Incentive") %>%
              group_by(year) %>%
              summarise(incent_perc = sum(FMV/year_tb_tot)),
            aes(x = year, y = incent_perc, color = "Incentive Classes"), lwd = 1 ) +
 
  # FMV with class 8 property class in cook county
   geom_line(data = class_8_df_outofCook %>%               # threw error here, missing x and y in aes()
              filter(class_1dig == 8) %>% 
              group_by(year) %>%
              summarize(perc_8 =  sum(FMV/year_tb_tot)),   ## added this part
            aes(x = year, y = perc_8, color = "Class 8"), lwd = 1) +        # was missing a + sign here
 
  # industrial fmv in cook county
  geom_line(data = class_8_df_outofCook %>%               # threw error here, missing x and y in aes()
              filter(Alea_cat == "Industrial") %>%
              group_by(year) %>%
              # needed na.rm=TRUE, otherwise it didn't work. perc_industrial was not being calculated without it.
              summarize(perc_industrial =  sum(FMV/year_tb_tot, na.rm=TRUE)),   ## added this part
            aes(x = year, y = perc_industrial, color = "Industrial"), lwd = 1) +

  # commercial fmv in cook county
geom_line(data = class_8_df_outofCook %>%               # threw error here, missing x and y in aes()
            filter(Alea_cat == "Commercial") %>%
            group_by(year) %>%
            summarize(perc_commercial =  sum(FMV/year_tb_tot, na.rm=TRUE)),   ## added this part
          aes(x = year, y = perc_commercial, color = "Commercial"), lwd = 1) +

  # make it pretty:
   theme_classic() +
  scale_x_continuous(name = "", breaks = seq(2006, 2022, by = 3), limits = c(2006, 2022), expand = c(0,0)) +
  scale_y_continuous(name = "Percent of County FMV", labels = scales::percent_format(),  limits = c(0, 0.20), 
                     breaks = seq(0, 0.5, by = 0.05), expand = c(0,0))  +
  scale_color_manual(name = "", values = c("Commercial+Industrial" = "black", "Industrial" = "gray80", "Commercial" = "gray40", "Incentive Classes" = "orange", "Class 8" =  "red")) +
  theme(legend.position = "bottom") +
  labs(title= "Cook County Commercial & Industrial FMV") + guides(color = guide_legend(nrow=2, byrow = TRUE))

## Newest Addition for April 30th Presentation --------------------
ggplot() +
  # incentive class properties in cook
  geom_line(data = class_8_df_outofCook %>%
              filter(incent_prop == "Incentive") %>%
              group_by(year) %>%
              summarise(incent_perc = sum(FMV/year_ind_comm_FMV)),
            aes(x = year, y = incent_perc, color = "Incentive Classes"), lwd = 1 ) +
  
  # FMV with class 8 property class in cook county
  geom_line(data = class_8_df_outofCook %>%               
              filter(class_1dig == 8) %>% 
              group_by(year) %>%
              summarize(perc_8 =  sum(FMV/year_ind_comm_FMV)),   
            aes(x = year, y = perc_8, color = "Class 8"), lwd = 1) +       
  # make it pretty:
  theme_classic() +
  scale_x_continuous(name = "", breaks = seq(2006, 2022, by = 3), limits = c(2006, 2022), expand = c(0,0)) +
  scale_y_continuous(name = "Percent of Com+Ind FMV", labels = scales::percent_format(), limits = c(0, 0.25), 
                     breaks = seq(0, 0.5, by = 0.05), expand = c(0,0))  +
  scale_color_manual(name = "", values = c("Industrial" = "gray80", "Commercial" = "gray40", "Incentive Classes" = "orange", "Class 8" =  "red")) +
  theme(legend.position = "bottom") +
  labs(title= "Cook County", subtitle =  "Share of Commercial & Industrial FMV with Incentive Classification") + guides(color = guide_legend(nrow=2, byrow = TRUE))


 ggplot() +
   geom_line(data = class_8_df %>%
               filter(incent_prop == "Incentive") %>%
               group_by(year) %>%
               summarise(incent_perc = sum(FMV/year_ind_comm_FMV)),
             aes(x = year, y = incent_perc, color = "All Incentive Classes"), lwd = 1 ) +
      geom_line(data = class_8_df %>%               
               filter(class_1dig == 8) %>% 
               group_by(year) %>%
               summarize(perc_8 =  sum(FMV/year_ind_comm_FMV)), 
             aes(x = year, y = perc_8, color = "Class 8"), lwd = 1) +        
   # make it pretty:
   theme_classic() +
   scale_x_continuous(name = "", breaks = seq(2006, 2022, by = 3), limits = c(2006, 2022), expand = c(0,0)) +
   scale_y_continuous(name = "Percent of Com&Ind FMV", labels = scales::percent_format(), # limits = c(0, 0.25), 
                      breaks = seq(0, 0.5, by = 0.05), expand = c(0,0))  +
   scale_color_manual(name = "", values = c("Commercial+Industrial" = "black", "Industrial" = "gray80", "Commercial" = "gray40", "All Incentive Classes" = "orange", "Class 8" =  "red")) +
   theme(legend.position = "bottom") +
   labs(title= "Class 8 Townships", subtitle =  "Share of Commercial & Industrial FMV with Incentive Classification") + guides(color = guide_legend(nrow=2, byrow = TRUE))
 
 
 
## Other variations --------------------------

## this one doesn't work because you never pass a data frame to it..

# Linegraph
# ggplot() + 
#   geom_line(aes(x = year, y = total_incent_ratio, color = "Ind. & Comm. FMV"), linewidth = 1.5) +
#   geom_line(aes(x = year, y = inc_ind_ratio, color = "Industrial %"), linewidth = 1.5) +
#   geom_line(aes(x = year, y = inc_com_ratio, color = "Commercial %"), linewidth = 1.5) +
#   geom_line(aes(x = year, y = class_8_ratio, color = "Class-8 %"), linewidth = 1.5) +
#   labs(y = "percent") +
#   scale_color_manual(values = c("#b2182b", "#fd8d3c", "#878787", "#000000")) +
#   theme_classic() +
#   scale_x_continuous(breaks = seq(2006, 2022, by = 3)) +
#   scale_y_continuous(labels = scales::percent_format(), limits = c(0, 0.45), breaks = seq(0, 0.5, by = 0.05)) +
#   guides(color = guide_legend(title = NULL))


plot_df <- muni_MC %>%
 # left_join(muni_MC, class_dict, by = "class") %>%
  mutate(FMV = av/assess_ratio) %>%
  select(year, class_1dig, incent_prop, Alea_cat, FMV) %>%
  filter(Alea_cat %in% c("Commercial", "Industrial")) %>%
  reframe(FMV = sum(FMV, na.rm=TRUE), .by = c("year", "class_1dig", "Alea_cat", "incent_prop")) %>%
  group_by(year) %>%
  mutate(year_sum_FMV = sum(FMV, na.rm = TRUE)) %>%
  ungroup() %>%
  group_by(year, Alea_cat) %>%
  mutate(cat_year_FMV = sum(FMV, na.rm=TRUE)) %>%
  ungroup() %>%
 # mutate(class_8 = ifelse(class_1dig == 8, 1, 0)) %>%
  group_by(year #, class_8
           ) %>%
  ## changed how year_class_8_FMV was calculated. sums if class is 8, otherwise leaves it as 0
  mutate(year_class_8_FMV = ifelse(class_1dig == 8, sum(FMV, na.rm=TRUE), 0)) %>%
  ungroup() %>%
  arrange(desc(year))

plot_df_FMV_cook <- muni_MC %>%
 # left_join(muni_MC, class_dict, by = "class") %>%
  mutate(FMV = av/assess_ratio) %>%
  select(year, class_1dig, incent_prop, Alea_cat, FMV) %>%
  filter(Alea_cat %in% c("Commercial", "Industrial")) %>%
  group_by(year) %>%
  summarize(year_max_tb = sum(FMV)) %>%
 # select(year, year_max_tb) %>%
 # distinct() %>%
  arrange(year)


# is this for commercial FMV totals for county or class 8 munis? 

plot_df_comm <-  muni_MC %>%
  filter(Alea_cat == "Commercial") %>%
  mutate(FMV = av/assess_ratio) %>%
 # select(year, class_1dig, incent_prop, Alea_cat, FMV) %>%
 # group_by(year, class_1dig, incent_prop, FMV) %>%
  reframe(year, class_1dig, incent_prop, Alea_cat, FMV,
          comm_tb = sum(FMV, na.rm=TRUE), .by = year) %>%
#  mutate(comm_tb = sum(FMV)) %>%    # total commercial FMV each year 
 filter(class_1dig %in% c(7, 8)) %>%   
  #distinct() %>%
  reframe(year, comm_tb,
          comm_inc_FMV = sum(FMV, na.rm=TRUE), .by = year) %>%
 # mutate(comm_inc_FMV = sum(FMV)) %>% # amount FMV incentivized each year
 # ungroup() %>%
  arrange(year) %>%
#  select(year, comm_inc_FMV, comm_tb) %>%   ### needed to select year here!!
  distinct()

plot_df_ind <-  muni_MC %>%
  #left_join(muni_MC, class_dict, by = "class") %>%
  mutate(FMV = av/assess_ratio) %>%
  select(year, class_1dig, incent_prop, Alea_cat, FMV) %>%
  filter(Alea_cat %in% c("Industrial")) %>%
  group_by(year) %>%
  mutate(ind_tb = sum(FMV)) %>%
  filter(class_1dig %in% c(6, 8)) %>%
  mutate(ind_inc_FMV = sum(FMV)) %>%
  arrange(year) %>%
  select(ind_inc_FMV, ind_tb) %>%
  distinct()

plot_df_8 <- muni_MC %>% 
  # left_join(muni_MC, class_dict, by = "class") %>%
  mutate(FMV = av/assess_ratio) %>%
  select(year, class_1dig, incent_prop, Alea_cat, FMV) %>%
  filter(Alea_cat %in% c("Commercial", "Industrial")) %>%
  group_by(year) %>%
  mutate(class_8 = ifelse(class_1dig == 8, 1, 0)) %>%
  filter(class_8 == 1) %>%
  group_by(year) %>%
  mutate(year_class_8_FMV = sum(FMV)) %>%
  arrange(year) %>%
  select(year_class_8_FMV) %>%
  distinct()

plot_df_final <- plot_df_FMV_cook %>%    # changed initial dataframe
  # plot_df_FMV %>%
  left_join(plot_df_8) %>%
  left_join(plot_df_comm) %>%
  left_join(plot_df_ind) %>%
#mutate ratio variables
  mutate(class_8_ratio = year_class_8_FMV/year_max_tb,
         inc_ind_ratio = ind_inc_FMV/ind_tb,
         inc_com_ratio = comm_inc_FMV/comm_tb,
         total_incent_ratio =
           (year_class_8_FMV + ind_inc_FMV + comm_inc_FMV)/year_max_tb)

#make line graph

plot_df_final %>%
  ggplot() +
  geom_line(aes(x = year, y = total_incent_ratio, color = "Ind. & Comm. FMV"), linewidth = 1.5) +
  geom_line(aes(x = year, y = inc_ind_ratio, color = "Industrial %"), linewidth = 1.5) +
  geom_line(aes(x = year, y = inc_com_ratio, color = "Commercial %"), linewidth = 1.5) +
  geom_line(aes(x = year, y = class_8_ratio, color = "Class-8 %"), linewidth = 1.5) +
  labs(y = "percent") +
  scale_color_manual(values = c("#b2182b", "#fd8d3c", "#878787", "#000000")) +
  theme_classic() +
  theme(legend.position = "bottom") + 
  scale_x_continuous(breaks = seq(2006, 2022, by = 3)) +
  scale_y_continuous(labels = scales::percent_format(), limits = c(0, 0.45), breaks = seq(0, 0.5, by = 0.05)) +
  guides(color = guide_legend(title = NULL), color = guide_legend(nrow=2, byrow = TRUE))


#### Fixed above dataframe and line graph ####### 

###### Stopped HERE night of 4/19 #### 

# Class 8 Muni Trends over Time

class8_df <- read_csv("./Output/class_8_ind_comm_FMV.csv")

class8_df_sums <- class8_df %>%
  group_by(year, incentive, type) %>%
  summarize(year, incentive, type, sum(FMV))


#make bar chart (NOT A HISTOGRAM)

## Michael version ## 
# plot_df_final %>%
#   ggplot() +
#   geom_col(aes(x = year - 0.1, y = total_incent_ratio, fill = "Total Incentives"),# width = .1, 
#            position = position_dodge(width = 0.2)) +
#   geom_col(aes(x = year, y = inc_ind_ratio, fill = "Industrial"), 
#            position = position_dodge(width = 0.2), width = .1) +
#   geom_col(aes(x = year + 0.1, y = inc_com_ratio, fill = "Commercial"), 
#            position = position_dodge(width = 0.2), width = .1) +
#   scale_fill_manual(values = c("Total" = "#fecc5c", "Industrial" = "#fd8d3c", "Commercial" = "#bd0026")) +
#   theme_classic() +
#   labs(x = "Year", y = "Ratio") + theme()


### Simple Version ###
plot_df_final %>%
  ggplot( ) +
  
  # this was too small to show up in the graph. It would need to become a separate graph. 
  geom_col(aes(x = year , y = total_incent_ratio, fill = "Total Incentives")) + # First I thought it wasn't working
  # then I realized that they are graphed on top of each other. Total incentives is hidden behind 
  #  commercial and industrial incentive bars.
  
  # geom_col(aes(x = year, y = inc_ind_ratio, fill = "Industrial")) + # commented out to see what f irst layer looked like
  # geom_col(aes(x = year, y = inc_com_ratio,fill = "Commercial")) +  # commented out to see what f irst layer looked like
  theme_classic() +
  labs(x = "Year", y = "FMV Ratio", title = "Incentive FMV / All FMV in Cook County")
  


plot_df_final %>%
  ggplot( ) +

  # Switch order of layers to get all 3 bars to appear!
  geom_col(aes(x = year, y = inc_ind_ratio, fill = "Industrial-Incentive")) +
  # this was too small to show up in the graph because it was behind other layers  
  geom_col(aes(x = year , y = total_incent_ratio, fill = "Incentive Classes / Total FMV")) + # First I thought it wasn't working
   geom_col(aes(x = year, y = inc_com_ratio, fill = "Commercial-Incentive")) +
  theme_classic() +
  labs(x = "Year", y = "FMV Ratio", title = "Industrial & Commercial \nFMV with Incentive Classification" ) 




plot_df_final %>%
  ggplot(
    ) +

  geom_col(aes(x = year, y = inc_ind_ratio,
               fill = "Industrial"), 
         #  position = position_dodge(width = 0.2), # width = .1
           ) +
  geom_col(aes(x = year, y = inc_com_ratio, 
               fill = "Commercial"), 
          # position = position_dodge(width = 0.2), # width = .1
           ) +
  #scale_fill_manual(values = c("Total" = "#fecc5c", "Industrial" = "#fd8d3c", "Commercial" = "#bd0026")) +
  theme_classic() +
  geom_line(aes(x = year , y = total_incent_ratio, 
                color = "Total Incentives"),# width = .1, 
            #position = position_dodge(width = 0.2)
  ) +
  labs(x = "Year", y = "% of FMV") 


plot_df_final %>%
  ggplot() +
  geom_col(aes(x = year - 0.15, y = total_incent_ratio, fill = "Total"), # width = .1,
           position = position_dodge(width = 0.2)) +
  geom_col(aes(x = year - 0.05, y = inc_ind_ratio, fill = "Industrial"),
           position = position_dodge(width = 0.2), 
        #   width = .1
           ) +
  geom_col(aes(x = year + 0.05, y = inc_com_ratio, fill = "Commercial"),
           position = position_dodge(width = 0.2), #width = .1
           ) +
  geom_col(aes(x = year + 0.15, y = class_8_ratio, fill = "Class-8"),
           position = position_dodge(width = 0.2), 
         #  width = .1
           ) +
  scale_fill_manual(values = c("Total" = "#fecc5c", "Industrial" = "#fd8d3c",
                               "Commercial" = "#bd0026", "Class-8" = "#000000")) +
  theme_classic() +
  labs(x = "Year", y = "Percent") +
  theme(legend.position = "bottom") + 
  scale_y_continuous(labels = scales::percent_format(), limits = c(0, 0.45), breaks = seq(0, 0.45, by = 0.05)) +
  scale_x_continuous(breaks = seq(2006, 2022, by = 3)) +
  guides(fill = guide_legend(title = NULL))




### Switch order of columns to make them show up better. Total was behind other layer.
plot_df_final %>%
  ggplot() +
  geom_col(aes(x = year - 0.15, y = total_incent_ratio, fill = "Total"), # width = .1,
           position = position_dodge(width = 0.2)) +
  geom_col(aes(x = year - 0.05, y = inc_ind_ratio, fill = "Industrial"),
           position = position_dodge(width = 0.2), 
           #   width = .1
  ) +
  
  geom_col(aes(x = year - 0.15, y = total_incent_ratio, fill = "Incentive FMV/Total FMV"), # width = .1,
           position = position_dodge(width = 0.2)) +
  
  
  geom_col(aes(x = year + 0.05, y = inc_com_ratio, fill = "Commercial"),
           position = position_dodge(width = 0.2), #width = .1
  ) +
  geom_col(aes(x = year + 0.15, y = class_8_ratio, fill = "Class-8"),
           position = position_dodge(width = 0.2), 
           #  width = .1
  ) +
  scale_fill_manual(values = c("Incentive FMV/Total FMV" = "#fecc5c", "Industrial" = "#fd8d3c",
                               "Commercial" = "#bd0026", "Class-8" = "#000000")) +
  theme_classic() +
  labs(x = "Year", y = "Percent") +
  theme(legend.position = "bottom", 
        legend.title = element_blank()) + 
  scale_y_continuous(labels = scales::percent_format(), limits = c(0, 0.45), breaks = seq(0, 0.45, by = 0.05)) +
  scale_x_continuous(breaks = seq(2006, 2022, by = 3)) 




# Scatterplot
## Goal: scatterplot where change in tax rate from exemptions on x-axis and change from incentives on y-axis
## Dot size represents total tax rate change
## Another color dot will represent municipalities

## Generate data

rates <- read_csv("alt_rates.csv")

delta_rates <- rates %>%
  select(clean_name, delta_exe = change_noExe, delta_inc = change_noInc, delta_both = change_neither) %>%
  distinct() %>%
  arrange(desc(delta_both)) %>%
  mutate(top_both = ifelse(row_number() %in% 1:20, 1, 0)
  ) %>%
  arrange(desc(delta_exe)) %>%
  mutate(top_exe = ifelse(row_number() %in% 1:20, 1, 0)) %>%
  arrange(desc(delta_inc)) %>%
  mutate(top_inc = ifelse(row_number() %in% 1:20, 1, 0
                          )) %>%
  mutate(key_obs = ifelse(top_exe == 1 & top_inc == 1, 1, 0)) %>%
  mutate(key_obs = as.factor(key_obs))

## Create plot

delta_rates %>%
  ggplot(aes(delta_exe, delta_inc)) +
  geom_point(aes(size = delta_both, color = key_obs), alpha = .4) +
  scale_size_continuous(guide = "none") +
  scale_color_manual(guide = "none", values = c("blue", "gray9")) +
  scale_x_continuous(expand = c(0,0), limits = c(0,16)) +
  scale_y_continuous(expand = c(0,0), limits = c(0,18)) +
  labs(
    x = "Exemptions",
    y = "Incentives",
    caption = "Axes represent percentage point change in the composite tax rate from 
    removing either incentive classifications or the General Homeowner Exemption. 
    Removing incentive classification implies taxing all commercial and industrial 
    property at at 25% level of assessment instead of 10% if it receives an incentive classification."
  ) +
  theme_classic()