Project 2025 Sentiment Analysis

Get Data from Project 2025 Website

Data collected from the Project 2025 website. Here is the python file used to collect and clean the data.

project2025 <- read_csv("https://jadenchant.github.io/data/project2025.csv") |>
  mutate(id = row_number()) |>
  select(id, section_name, chapter, chapter_name, subsection_name, authors, text) |>
  mutate(word_count = str_count(text, "\\S+"))

Bing Sentiment Analysis

bing_sentiment <- get_sentiments("bing")

tokenized <- project2025 |>
  mutate(text = as.character(text)) |>
  unnest_tokens(word, text)

bing_sentiment_count <- tokenized |>
  inner_join(bing_sentiment, by = "word") |>
  group_by(id, sentiment) |>
  summarize(count = n()) |>
  pivot_wider(names_from = sentiment, values_from = count, values_fill = 0) |>
  rename(bing_positive = positive, bing_negative = negative)
  
bing_result <- left_join(project2025, bing_sentiment_count, by = "id") |>
  mutate(bing_pred_sentiment = as.factor(ifelse(bing_positive > bing_negative, "positive", "negative"))) |>
  mutate(bing_pred_val = bing_positive / (bing_positive + bing_negative))

Afinn Sentiment Analysis

# afinn_sentiment <- get_sentiments("afinn")
# 
# afinn_sentiment_count <- tokenized |>
#   inner_join(afinn_sentiment, by = "word") |>
#   group_by(id) |>
#   summarize(
#     afinn_positive = sum(value[value > 0], na.rm = TRUE),
#     afinn_negative = sum(abs(value[value < 0]), na.rm = TRUE),
#     afinn_total = sum(value, na.rm = TRUE))
# 
# result <- left_join(project2025, afinn_sentiment_count, by = "id") |>
#   mutate(afinn_pred_sentiment = as.factor(ifelse(afinn_positive > afinn_negative, "positive", "negative"))) |>
#   mutate(afinn_pred_val = afinn_positive / (afinn_positive + afinn_negative))

result_long <- bing_result |>
  pivot_longer(cols = c(bing_positive, bing_negative), 
               names_to = "Sentiment", values_to = "Count") |>
  mutate(Sentiment = ifelse(Sentiment == "bing_positive", "Positive", "Negative"))

ggplot(result_long, aes(x = id, y = Count, color = Sentiment)) +
  geom_line() +
  labs(title = "Sentiment Analysis Throughout the Playbook", 
       subtitle = "(Bing Lexicon)",
       x = "Project 2025 Playbook", 
       y = "Sentiment Count", 
       color = "Sentiment Type") +
  scale_color_manual(values = c("Positive" = "#0a8754", "Negative" = "#AA0E3D")) +
  dark_theme_gray(base_size = 12) +
  theme(axis.text.x = element_blank(),
        axis.ticks.x = element_blank())

ggplot(result_long, aes(x = id, y = bing_pred_val, color = bing_pred_val)) +
  geom_line() +
  scale_color_gradient2(low = "#AA0E3D", high = "#0a8754", 
                        midpoint = 0.5, name = "Sentiment Percentage") +
  labs(title = "Sentiment Analysis Throughout the Playbook", 
       subtitle = "(Bing Lexicon)",
       x = "Project 2025 Playbook", 
       y = "Sentiment Percentage") +
  dark_theme_gray(base_size = 12) +
  theme(axis.text.x = element_blank(),
        axis.ticks.x = element_blank())