library(ggplot2)
library(dplyr)
library(tidyr)
library(readr)
library(stringr)
library(cowplot)
library(extrafont)
font_import(pattern = "Times", prompt = FALSE)

# This file contains the code to produce the figures and tables in the main body of the paper.

# Author of figure code: Christina Maimone christina.maimone@northwestern.edu


# Load data ----
patents <- read.csv("patents_analysis.csv")


# styling
basetheme <- theme(
  panel.grid.minor = element_blank(),
  panel.grid.major = element_blank(),
  panel.background = element_rect(fill = "transparent", color = "gray", size = .5),
  plot.background = element_rect(fill = "transparent", color = NA),
  axis.line.x = element_line(colour = "#666666", size = .25),
  axis.line.y = element_line(colour = "#666666", size =  .25),
  axis.ticks = element_line(colour = "#666666", size = .25),
  axis.text = element_text(color = "#333333", size = 9),
  legend.key = element_rect(fill = NA),
  legend.title = element_blank(),
  strip.background = element_blank(),
  strip.placement = "outside",
  legend.position = "none",
  text = element_text(family = "Times New Roman"),
  panel.spacing = unit(1.5, "lines"),
  axis.title = element_text(size = 9),
  plot.caption = element_text(size = 9)
)

# Figure 1  ----

figdata <- patents %>%
  select(validity,
         starts_with("forward_cite_year"),
         -ends_with("examiner")) %>%
  gather(key = "year",
         value = "citations",
         starts_with("forward_cite_year")) %>%
  mutate(year = parse_number(year),
         Validity = tools::toTitleCase(as.character(validity))) %>%
  mutate(Validity = factor(Validity, levels = c("Valid", "Invalid"))) %>%
  group_by(Validity, year) %>%
  summarize(mean = mean(citations))

xticklabels <- rep("",21)
xticklabels[seq(0,20,5)+1] <- seq(0,20,5)

f1 <- figdata %>%
  ggplot(aes(
    x = year,
    y = mean,
    color = Validity,
    shape = Validity
  )) +
  geom_point(size = 2) +
  geom_line() +
  scale_shape_manual(values = c(16, 15)) +
  labs(title = "",
    x = "Patent Term",
    y = "Average Forward Citations\nReceived in Year") +
  scale_y_continuous(limits = c(0, 12),
                     breaks=0:12,
                     labels = c(rbind(seq(0, 12, 2), ""))[-14]) +
  scale_x_continuous(limits = c(0, 20),
                     breaks=0:20,
                     labels=xticklabels) +
  scale_color_grey(start = .2, end = .6) +
  annotate(
    "text",
    x = c(15.5, 18.5),
    y = c(1.3, 4.8),
    label = c("Invalid", "Valid") ,
    color = "black",
    size = 3,
    family = "Times New Roman",
    hjust = 0
  ) +
  basetheme 

f1

ggsave("figure1.pdf", 
       plot = f1, 
       width = 4.67, height = 2.5)

ggsave("figure1.png", 
       plot = f1, 
       width = 4.67, height = 2.5)

f1 + 
  scale_y_continuous(limits = c(0, 12),
                     breaks=0:12,
                     labels = c(rbind(seq(0, 12, 2), ""))[-14],
                     expand=c(0,0,0,.1)) +
  theme(panel.grid.major.y = element_line(colour = "#cccccc", size = .25))

ggsave("figure1_grid.pdf", 
       width = 4.67, height = 2.5)
ggsave("figure1_grid.png", 
       width = 4.67, height = 2.5)

# Figure 2 ----

# Note: 18 data points are excluded from the plot because of the limits on the x axis

figdata2 <- patents %>%
  select(patent_num, validity, starts_with("forward_cite_year")) %>%
  gather(key = "var",
         value = "citations",
         starts_with("forward_cite_year")) %>%
  mutate(year = parse_number(var),
         var = ifelse(str_detect(var, "us"), "total", "examiner")) %>%
  spread(var, citations, fill = 0) %>%
  mutate(Validity = factor(
    tools::toTitleCase(as.character(validity)),
    levels = c("Valid", "Invalid")
  )) %>% # for plot capitalization
  group_by(Validity, year) %>%
  summarize(
    `Examiner Citations` = mean(examiner, na.rm = TRUE),
    `Applicant Citations` = mean(total - examiner, na.rm = TRUE)
  ) %>%
  gather(key = "type", value = "mean", contains("Citations"))


f2 <- figdata2 %>%
  ggplot(aes(
    x = year,
    y = mean,
    color = Validity,
    shape = Validity
  )) +
  geom_point(size = 2) +
  facet_wrap(type ~ ., ncol = 1, scales = "free") +
  geom_line() +
  scale_shape_manual(values = c(16, 15)) +
  labs(title = "",
    x = "Patent Term",
    y = "Average Forward Citations Received in Year") +
  scale_y_continuous(limits = c(0, 7), breaks = seq(0, 7, 1)) +
  scale_x_continuous(limits = c(0, 20),
                     breaks=0:20,
                     labels=xticklabels) +
  scale_color_grey(start = .2, end = .6) +
  basetheme +
  theme(
    strip.text = element_text(size = 9, margin = margin(b = 2)),
    axis.title.y = element_text(margin = margin(r = 10))
  )

ggdraw(f2) +
  draw_label(
    "Invalid", x = .89, y = .62, vjust = 1, hjust = 1, size = 8, fontfamily = "Times New Roman"
  ) +
  draw_label(
    "Valid", x = .96, y = .70, vjust = 1, hjust = 1, size = 8, fontfamily = "Times New Roman"
  ) +
  draw_label(
    "Invalid", x = .77, y = .155, vjust = 1, hjust = 1, size = 8, fontfamily = "Times New Roman"
  ) +
  draw_label(
    "Valid", x = .91, y = .24, vjust = 1, hjust = 1, size = 8, fontfamily = "Times New Roman"
  )

ggsave("figure2.png", width = 4.67, height = 5)
ggsave("figure2.pdf", width = 4.67, height = 5)


f3 <- f2 + 
  scale_y_continuous(limits = c(0, 7), breaks = seq(0, 7, 1),
                     expand=c(0,0,0,0.1)) +
  theme(panel.grid.major.y = element_line(colour = "#cccccc", size = .25),
        strip.text= element_text(margin=margin(b=.5, unit = "cm")))

ggdraw(f3) +
  draw_label(
    "Invalid", x = .89, y = .60, vjust = 1, hjust = 1, size = 8, fontfamily = "Times New Roman"
  ) +
  draw_label(
    "Valid", x = .96, y = .685, vjust = 1, hjust = 1, size = 8, fontfamily = "Times New Roman"
  ) +
  draw_label(
    "Invalid", x = .77, y = .14, vjust = 1, hjust = 1, size = 8, fontfamily = "Times New Roman"
  ) +
  draw_label(
    "Valid", x = .9, y = .225, vjust = 1, hjust = 1, size = 8, fontfamily = "Times New Roman"
  )

ggsave("figure2_grid.png", width = 4.67, height = 5)
ggsave("figure2_grid.pdf", width = 4.67, height = 5)



# Figure 3 ----

xticklabels <- rep("",17)
xticklabels[c(1,5,9,13,17)] <- seq(0,800,200)

patents %>%
  mutate(Validity = tools::toTitleCase(as.character(validity))) %>%
  mutate(Validity = factor(Validity, levels = c("Invalid", "Valid"))) %>%
  mutate(tech_class_6 = str_wrap(tech_class_6, 12)) %>%
  ggplot(aes(y = forward_cite_us, x = Validity, fill = Validity)) +
  geom_boxplot(
    outlier.size = 1,
    outlier.alpha = .5,
    lwd = .25,
    outlier.stroke = 0
  ) +
  coord_flip() +
  facet_grid(tech_class_6 ~ ., switch = "both", scales = "free") +
  scale_y_continuous(limits = c(0, 800), expand = c(0, 0),
                     breaks=seq(0,800,50),
                     labels=xticklabels) +
  scale_fill_grey(start = .9, end = .65) +
  labs(
    title = "",
    y = "Forward Citations*",
    caption = "*Range shown limited to 800",
    x = ""
  ) +
  basetheme +
  theme(
    strip.text = element_text(
      vjust = 0,
      margin = margin(0, .25, 0, 0, "cm"),
      size = 9
    ),
    axis.text.y = element_text(size = 9),
    axis.ticks.y = element_blank(),
    plot.margin = unit(c(-.3, .3, 0,-.2), "cm"),
    panel.spacing = unit(1, "lines"),
    plot.caption = element_text(size=8)
  )

ggsave("figure3.png", width = 4.67, height = 5)
ggsave("figure3.pdf", width = 4.67, height = 5)


# Figure 3 summary stats and t-tests

tbl <- patents %>%
  group_by(tech_class_6, validity) %>%
  summarize(
    Count = n(),
    `Median Forward Citations` = median(forward_cite_us),
    `Mean Forward Citations` = mean(forward_cite_us),
    `Mean log(Forward Citations+1)` = mean(log(1 + forward_cite_us))
  ) %>%
  mutate(`Difference in Logged Means (Valid-Invalid)` = `Mean log(Forward Citations+1)`[2] -
           `Mean log(Forward Citations+1)`[1])
tbl[seq(2, 12, 2), "Difference in Logged Means (Valid-Invalid)"] <-
  NA
tbl[1, "p-value on t-test of Difference"] <-
  t.test(log(forward_cite_us + 1) ~ validity,
         subset = tech_class_6 ==
           "Chemical",
         data = patents)$p.value
tbl[3, "p-value on t-test of Difference"] <-
  t.test(
    log(forward_cite_us + 1) ~ validity,
    subset = tech_class_6 ==
      "Computers & Communication",
    data = patents
  )$p.value
tbl[5, "p-value on t-test of Difference"] <-
  t.test(
    log(forward_cite_us + 1) ~ validity,
    subset = tech_class_6 ==
      "Drugs & Medical",
    data = patents
  )$p.value
tbl[7, "p-value on t-test of Difference"] <-
  t.test(
    log(forward_cite_us + 1) ~ validity,
    subset = tech_class_6 ==
      "Electrical & Electronics",
    data = patents
  )$p.value
tbl[9, "p-value on t-test of Difference"] <-
  t.test(
    log(forward_cite_us + 1) ~ validity,
    subset = tech_class_6 ==
      "Mechanical",
    data = patents
  )$p.value
tbl[11, "p-value on t-test of Difference"] <-
  t.test(log(forward_cite_us + 1) ~ validity,
         subset = tech_class_6 ==
           "Others",
         data = patents)$p.value
tbl



# Figure 4 ----

# Note: 24 data points are excluded from the plot because of the limits on the x axis


tmp <- patents %>%
  select(forward_cite_us,
         invalidation_s102,
         invalidation_s103,
         invalidation_s112) %>%
  gather("validity",
         "value",
         invalidation_s102,
         invalidation_s103,
         invalidation_s112) %>%
  filter(value == 1) %>%
  select(-value) %>%
  mutate(validity = paste0("35 U.S.C. § ", parse_number(validity)))

fig4data <- patents %>%
  select(forward_cite_us, validity) %>%
  mutate(validity = ifelse(validity == "valid", "All Valid", "All Invalid")) %>%
  rbind(tmp)

fig4data <- fig4data %>%
  mutate(validity = factor(
    validity,
    levels = c(
      "All Valid",
      "All Invalid",
      "35 U.S.C. § 102",
      "35 U.S.C. § 103",
      "35 U.S.C. § 112"
    )
  )) %>%
  mutate(forward_cite_us_log = log(forward_cite_us + 1))

fig4data %>%
  mutate(validity = factor(validity, levels = rev(levels(validity)))) %>%
  ggplot(aes(y = forward_cite_us, x = validity)) +
  geom_boxplot(
    outlier.size = 1,
    outlier.alpha = .5,
    lwd = .25,
    outlier.stroke = 0,
    fill = c(gray(.9), gray(.9), gray(.9), gray(.9), gray(.65)),
    width = .6
  ) +
  coord_flip() +
  scale_y_continuous(limits = c(0, 800), expand = c(0, 0),
                     breaks=seq(0,800,50),
                     labels=xticklabels) +
  scale_fill_grey(start = .9, end = .65) +
  labs(
    title = "",
    x = "",
    y = "Forward Citations*",
    caption = "*Range shown limited to 800"
  ) +
  basetheme +
  theme(#axis.line.y=element_blank(),
    axis.ticks.y = element_blank(),
    plot.margin = unit(c(-.3, .3, 0,-.2), "cm"),
    plot.caption = element_text(size=8))

ggsave("figure4.png", width = 4.67, height = 2)
ggsave("figure4.pdf", width = 4.67, height = 2)


# Figure 4 t-tests

p <- c(0, 0, 0, 0)
names(p) <- levels(fig4data$validity)[-1]
p[1] <-
  t.test(fig4data$forward_cite_us_log[fig4data$validity == "All Valid"],
         fig4data$forward_cite_us_log[fig4data$validity == "All Invalid"])$p.value
p[2] <-
  t.test(fig4data$forward_cite_us_log[fig4data$validity == "All Valid"],
         fig4data$forward_cite_us_log[fig4data$validity == "35 U.S.C. § 102"])$p.value
p[3] <-
  t.test(fig4data$forward_cite_us_log[fig4data$validity == "All Valid"],
         fig4data$forward_cite_us_log[fig4data$validity == "35 U.S.C. § 103"])$p.value
p[4] <-
  t.test(fig4data$forward_cite_us_log[fig4data$validity == "All Valid"],
         fig4data$forward_cite_us_log[fig4data$validity == "35 U.S.C. § 112"])$p.value
p <- round(p, 2)
p <- ifelse(p < .001, "< 0.001", p)

test_tbl <- fig4data %>%
  rename(Validity = validity) %>%
  group_by(Validity) %>%
  summarize(Count = n(),
            `Mean (log forward citations)` = mean(log(forward_cite_us + 1))) %>%
  mutate(p = c(NA, p))
test_tbl