BCB 520 - Midterm Portfolio Post

Analysis of University of Idaho’s Research Funding Performance

This post analyzes the University of Idaho’s research funding portfolio, including active awards, proportional representation of new awards, and performance compared to peer institutions. The analysis uses data visualization and justified metrics to provide insights for university leadership.
research funding
data visualization
university comparison
Author

Yaotian Gao

Published

March 20, 2024

1. Preamble

The University of Idaho (UI) relies on research funding from various sponsors to support its academic mission and drive innovation. Understanding the current state and future outlook of the university’s research funding portfolio is crucial for informed decision-making by university leadership. This post aims to address three key questions:

  1. What is the current state of UI’s active research awards, including their duration, amount, and principal investigators?
  2. How has the proportional representation of new awards from different funding sources changed over the past 5 to 10 years, and are there any notable trends?
  3. How does UI’s performance in securing research funding compare to that of peer institutions?

By analyzing data related to these questions, this post will provide valuable insights into UI’s research funding landscape. The findings will help university leadership understand the institution’s strengths, identify areas for improvement, and make data-driven decisions to support the university’s research goals. Furthermore, this analysis will contribute to the university’s strategic planning efforts and help ensure the sustainability and growth of its research enterprise.

2. Data and Code

Data Sources Summary: This analysis utilizes data from four major U.S. research funding agencies: the Department of Agriculture (USDA NIFA), Department of Energy(DOE), National Institutes of Health (NIH), and National Science Foundation (NSF). The data was obtained from the following sources:

Data Source Description URL
NIFA Award Database Public database of awards granted by the NIFA https://portal.nifa.usda.gov/lmd4/recent_awards
Department of Energy PAMS External award search interface for the Department of Energy’s PAMS https://pamspublic.science.energy.gov/WebPAMSExternal/interface/awards/AwardSearchExternal.aspx
NIH RePORTER API API providing access to NIH research grant data https://api.reporter.nih.gov/documents/Data%20Elements%20for%20RePORTER%20Project%20API_V2.pdf
NSF Awards API API offering access to NSF award data https://resources.research.gov/common/webapi/awardapisearch-v1.html

This data dictionary table provides an overview of each data source used in the analysis, including a brief description and the URL where the data can be accessed. The combination of these four data sources allows for a comprehensive analysis of research funding from major U.S. agencies to the University of Idaho and its peer institutions.

The raw data and code used in this analysis are provided in the supplementary materials.

3. Results

3.1 Visualizing the University of Idaho’s Active Research Awards by Sponsor

To address Question 1, we investigated the funding situation of three major sponsors (DOE, NIH, and NSF) for the University of Idaho, focusing on projects with end dates in 2024 and beyond. Due to the lack of end date information provided by USDA, this analysis does not include data from USDA.

We utilized two types of bar charts to present the findings. Figure 1 is a simple bar chart that arranges sponsors in order of increasing number of projects expiring in 2024. Projects expiring in 2024 are highlighted using a red channel to emphasize the expiration timeline for each sponsor. Figures 2-4 use stacked bar charts, with principal investigators (PIs) on the x-axis and the total amount of funding for projects ending in 2024 and beyond on the y-axis. Similarly, projects expiring in 2024 are represented using a red channel. The PIs are ordered by decreasing total funding amount, allowing for quick identification of the researchers with the most funding.

Examining each sponsor individually, it is evident that for both NIH and NSF, the top five PIs in terms of total funding received have significantly higher amounts compared to other researchers, displaying a clear “head” trend in funding distribution (Fig2, 4, Table S1).

3.3 Comparative Analysis of University of Idaho’s Performance with Peer Institutions

3.3.1 Data Preparation and Metrics

To assess the performance of each university, I selected data from the past five years (2019-2023) and processed the data as follows:

Calculated the average annual funding per project (Mean Award Amount) for each university from different sponsors. This was done by dividing the total funding from each sponsor in a given year by the number of projects. This approach considers both the total funding and the number of projects, eliminating differences due to university size. Subtracted the University of Idaho’s average project funding from each university’s average project funding to determine the relative funding difference. A positive value indicates that the other university received more funding than the University of Idaho, while a negative value indicates less funding. Added the relative funding differences from all four sponsors to obtain the overall funding difference between each university and the University of Idaho for each year (Fig 10). Used a bar chart to visually compare the differences between universities and the University of Idaho across different years, with colors distinguishing the years.

3.3.2 Analysis for each sponser

Overall (Fig 10), the University of Idaho(UI) performs better than Idaho State University, on par with Boise State University and the University of Montana, slightly worse than Washington State University, and significantly behind Montana State University.

3.3.2.1 DOE Analysis

The University of Idaho performs similarly to Idaho State University and the University of Montana, worse than Washington State University, and significantly behind Montana State University. Notably, Boise State University received a large grant in 2023, leading to its superior performance compared to UI (Fig 6).

Considering these results along with the findings from Q2 regarding the number of projects, UI’s competitiveness in DOE projects appears to be relatively weak.

3.3.2.2 NIH Analysis

UI performs better than Idaho State University but lags behind the other universities in all years except 2022. Combined with the declining trend in the number of NIH projects from Q2, it can be concluded that UI is less competitive than the other universities in securing NIH funding (Fig 7).

3.3.2.4 NSF Analysis

In NSF projects, UI consistently outperforms the other universities and maintains its lead. Considering the stable trend in the number of new NSF projects from Q2, UI appears to have a strong and consistent competitive edge in NSF projects (Fig 8).

3.3.2.5 USDA Analysis

For USDA projects, UI consistently outperforms Idaho State University, slightly surpasses Boise State University, marginally lags behind the University of Montana, and significantly trails Washington State University and Montana State University (Fig 9).

Overall, UI has some competitiveness in USDA projects but falls behind Montana State University and Washington State University.

Summary

This analysis aimed to provide a comprehensive understanding of the University of Idaho’s (UI) research funding portfolio by addressing three key questions: (1) the current state of active research awards, (2) trends in the proportional representation of new awards from various funding sources, and (3) UI’s performance compared to peer institutions.

The investigation of active research awards revealed that for NIH and NSF, the top five principal investigators (PIs) receive significantly higher funding amounts compared to other researchers, indicating a “head” trend in funding distribution. The analysis of new award trends over the past decade showed that USDA and NIH experienced overall declining trends, with USDA exhibiting more fluctuations and NIH demonstrating a more consistent decrease. In contrast, NSF maintained a relatively stable number of new projects. These findings suggest that UI may need to explore new funding sources to compensate for the declines in USDA and NIH funding while leveraging the stability of NSF support.

The comparative analysis of UI’s performance against peer institutions employed a metric based on the average annual funding per project, which considered both total funding and the number of projects. Overall, UI performed better than Idaho State University, on par with Boise State University and the University of Montana, slightly worse than Washington State University, and significantly behind Montana State University. UI’s competitiveness varied across sponsors, with a relatively weak position in DOE projects, less competitiveness in NIH funding compared to most peers, a strong and consistent competitive edge in NSF projects, and some competitiveness in USDA projects, although falling behind Montana State University and Washington State University.

These findings provide valuable insights into UI’s research funding landscape, highlighting strengths, areas for improvement, and potential strategies for ensuring the sustainability and growth of its research enterprise. The analysis emphasizes the importance of diversifying funding sources, particularly in light of declining trends from certain sponsors, and focusing on maintaining and enhancing competitiveness in areas of strength, such as NSF projects. Furthermore, the comparative analysis with peer institutions offers a benchmark for evaluating UI’s performance and identifies potential areas for learning and collaboration.

Supplementary Materials

Tabel S1: Active Research Awards by Sponsor, Principal Investigator, and End Date

Instrction Start year End year Award amount PI
DOE 2020 2024 603903 Yager, Elowyn
DOE 2021 2024 1404162 Marx, Christopher
DOE 2021 2024 1519359 Vasdekis, Andreas
DOE 2021 2024 1812000 Sammarruca, Francesca
NIH 2019 2024 61038 JOHNSON, JILL L
NIH 2021 2024 181821 GRIESHABER, SCOTT S
NIH 2019 2024 285011 JOHNSON, JILL L
NIH 2019 2024 346779 MARTIN, BRYN ANDREW
NIH 2018 2024 362003 FORTUNATO, ELIZABETH A
NIH 2018 2024 428581 MCGUIRE, MICHELLE KAY
NIH 2018 2024 460341 LUCKHART, SHIRLEY
NIH 2023 2025 34078 PATEL, JAGDISH SURESH
NIH 2023 2025 64727 VAN LEUVEN, JAMES
NIH 2015 2025 147472 FU, AUDREY QIUYAN
NIH 2015 2025 152932 VAN LEUVEN, JAMES
NIH 2015 2025 153149 FU, AUDREY QIUYAN
NIH 2015 2025 156572 PATEL, JAGDISH SURESH
NIH 2015 2025 163515 REMIEN, CHRISTOPHER HASKELL
NIH 2015 2025 181902 XIAN, MIN
NIH 2015 2025 199138 XIAN, MIN
NIH 2015 2025 266181 WICHMAN, HOLLY A
NIH 2020 2025 348389 MITCHELL, DIANA
NIH 2020 2025 357664 MITCHELL, DIANA
NIH 2020 2025 359164 MITCHELL, DIANA
NIH 2020 2025 359906 MITCHELL, DIANA
NIH 2015 2025 492598 WICHMAN, HOLLY A
NIH 2015 2025 677022 MILLER, CRAIG R
NIH 2015 2025 715546 MILLER, CRAIG R
NIH 2015 2025 764637 MILLER, CRAIG R
NIH 2015 2025 1025023 WICHMAN, HOLLY A
NIH 2015 2025 1040007 WICHMAN, HOLLY A
NIH 2015 2025 1138359 WICHMAN, HOLLY A
NIH 2015 2025 2172986 WICHMAN, HOLLY A
NIH 2015 2025 2194934 WICHMAN, HOLLY A
NIH 2015 2025 2206645 WICHMAN, HOLLY A
NIH 2015 2025 2212500 WICHMAN, HOLLY A
NIH 2023 2026 171789 HERNANDEZ VARGAS, ESTEBAN ABERLARDO
NIH 2022 2026 544548 LUCKHART, SHIRLEY
NIH 2022 2026 556756 LUCKHART, SHIRLEY
NIH 2016 2027 367126 NUISMER, SCOTT L
NIH 2023 2028 98616 WAYNANT, KRISTOPHER V
NIH 2023 2028 263413 ROBISON, BARRIE D
NIH 2024 2029 153322 LANE, GINNY
NIH 2024 2029 162320 CHEN, YIMIN
NIH 2024 2029 300000 MCGUIRE, MICHELLE KAY
NIH 2024 2029 331232 BROWN, ANN FROST
NIH 2024 2029 461621 WILLIAMS, JANET E.
NIH 2024 2029 946131 MCGUIRE, MICHELLE KAY
NIH 2024 2029 2354626 MCGUIRE, MICHELLE KAY
NSF 2021 2024 78630 Erika Rader
NSF 2023 2024 99445 Indrajit Charit
NSF 2019 2024 108245 Leslie L Baker
NSF 2020 2024 117607 Grant Harley
NSF 2022 2024 118517 Zachariah B Etienne
NSF 2021 2024 128702 Grant Harley
NSF 2022 2024 226347 Zachariah B Etienne
NSF 2021 2024 227519 Karla Eitel
NSF 2021 2024 335902 Zachariah B Etienne
NSF 2022 2024 344599 Jessica R Stanley
NSF 2019 2024 348957 Elowyn Yager
NSF 2019 2024 500000 Andrew S Nelson
NSF 2021 2024 539714 William J Reeder
NSF 2016 2024 558035 Daniele Tonina
NSF 2020 2024 586223 Adam G Jones
NSF 2014 2024 597487 Jerry R McMurtry
NSF 2019 2024 651698 Michael S Strickland
NSF 2020 2024 671486 Timothy C Bartholomaus
NSF 2019 2024 729932 Elizabeth J Cassel
NSF 2018 2024 1777713 Christopher J Marx
NSF 2020 2024 5830709 Xiaogang Ma
NSF 2018 2024 20000000 Andrew D Kliskey
NSF 2023 2025 45380 Alexander Woo
NSF 2022 2025 140899 Eva M Top
NSF 2022 2025 199946 Johnny Li
NSF 2023 2025 208292 Elowyn Yager
NSF 2023 2025 250000 Esteban A Hernandez Vargas
NSF 2021 2025 336514 Andreas E Vasdekis
NSF 2022 2025 402341 Nathan R Schiele
NSF 2021 2025 443170 Leda N Kobziar
NSF 2019 2025 449355 Lilian Alessa
NSF 2018 2025 599643 Eric L Mittelstaedt
NSF 2021 2025 677575 Laurel Lynch
NSF 2019 2025 749639 Andrew D Kliskey
NSF 2021 2025 792475 Xiaogang Ma
NSF 2024 2025 1000000 Tara Hudiburg
NSF 2020 2025 1028242 Adam G Jones
NSF 2018 2025 1093880 Christine E Parent
NSF 2020 2025 1368804 Julie M Amador
NSF 2021 2025 3974309 Michael Maughan
NSF 2023 2026 156022 Grant Harley
NSF 2023 2026 359828 Amin Mirkouei
NSF 2023 2026 443525 Andrew Tranmer
NSF 2021 2026 499866 Christine E Parent
NSF 2023 2026 509641 Eric L Mittelstaedt
NSF 2024 2026 628415 Dave Lien
NSF 2023 2026 645273 Kristopher V Waynant
NSF 2023 2026 664468 Scott L Nuismer
NSF 2022 2026 2999999 Vanessa Anthony-Stevens
NSF 2021 2026 18950955 Michael S Strickland
NSF 2022 2027 276000 Jerry R McMurtry
NSF 2024 2027 456051 Kristopher V Waynant
NSF 2022 2027 898220 Paul A Rowley
NSF 2022 2027 1016312 Chris A Hamilton
NSF 2023 2027 1179977 Julie M Amador
NSF 2023 2027 2435509 Lilian Alessa
NSF 2022 2027 4460195 Jim Alves-Foss
NSF 2023 2028 250000 Shirley Luckhart
NSF 2023 2028 20000000 Andrew D Kliskey

Code Fold

Q1

Code 
# Scale drawing
library(ggplot2)

# Create data frame
data <- data.frame(
  Sponsor = c("DOE", "NSF", "NIH"),
  Expiring_2024 = c(4, 22, 7),
  Active_After_2024 = c(0, 37, 38)
)

# Converts the data frame to a long format
data_long <- tidyr::pivot_longer(data, cols = c("Expiring_2024", "Active_After_2024"), 
                                 names_to = "Status", values_to = "Count")

# Draw a group bar chart
ggplot(data_long, aes(x = Sponsor, y = Count, fill = Status)) + 
  geom_col(position = "dodge") +
  geom_text(aes(label = Count), position = position_dodge(width = 0.9), 
            vjust = -0.5, size = 4) +
  scale_fill_manual(values = c("Expiring_2024" = "red", "Active_After_2024" = "blue"),
                    labels = c("Active After 2024", "Expiring in 2024")) +
  labs(title = "Fig1. Project Status by Sponsor",
       x = "Sponsor", y = "Number of Projects") +
  theme_minimal() +
  theme(plot.title = element_text(hjust = 0.5, face = "bold"))


#NSF

library(ggplot2)
library(dplyr)
# Read data

data <- read.csv("NSF-UI.csv", stringsAsFactors = FALSE)

# Screening projects for 2024 and beyond
data_filtered <- data %>% 
  filter(End.year >= 2024) %>%
  select(End.year, Award.amount, PI)
  
# Replace years After 2024 with "After 2024"
data_filtered$End.year <- ifelse(data_filtered$End.year > 2024, "After 2024", as.character(data_filtered$End.year))


# Sorted in descending order by Award.amount
data_filtered <- data_filtered %>%
  arrange(desc(Award.amount))

# Convert PI to a factor variable and arrange the levels in order of Award.amount
data_filtered$PI <- factor(data_filtered$PI, levels = unique(data_filtered$PI))

# Draw a column chart
ggplot(data_filtered, aes(x = PI, y = Award.amount, fill = factor(End.year))) +
  geom_col() +
  #geom_text(aes(label = Award.amount), vjust = -0.5, size = 3) +
  scale_fill_manual(values = c("2024" = "red", "After 2024"="blue"), name = "End Year") +
  labs(title = "Fig2.NSF Awards to University of Idaho (2024 and Beyond)",
       x = "Principal Investigator", y = "Award Amount") +
  theme_minimal() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1),
        legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
        
#DOE
library(ggplot2)
library(dplyr)
# Read data

data <- read.csv("DOE-UI.csv", stringsAsFactors = FALSE)

# Screening projects for 2024 and beyond
data_filtered <- data %>% 
  filter(End.year >= 2024) %>%
  select(End.year, Award.amount, PI)
  
# Replace years After 2024 with "After 2024"
data_filtered$End.year <- ifelse(data_filtered$End.year > 2024, "After 2024", as.character(data_filtered$End.year))


# Sorted in descending order by Award.amount
data_filtered <- data_filtered %>%
  arrange(desc(Award.amount))

# Convert PI to a factor variable and arrange the levels in order of Award.amount
data_filtered$PI <- factor(data_filtered$PI, levels = unique(data_filtered$PI))

# Draw a column chart
ggplot(data_filtered, aes(x = PI, y = Award.amount, fill = factor(End.year))) +
  geom_col() +
  #geom_text(aes(label = Award.amount), vjust = -0.5, size = 3) +
  scale_fill_manual(values = c("2024" = "red", "After 2024"="blue"), name = "End Year") +
  labs(title = "Fig3. DOE Awards to University of Idaho (2024 and Beyond)",
       x = "Principal Investigator", y = "Award Amount") +
  theme_minimal() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1),
        legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold")) 
        
#NIH
library(ggplot2)
library(dplyr)
# Draw a column chart

data <- read.csv("NIH-UI.csv", stringsAsFactors = FALSE)

# Filter projects from 2024 and onwards
data_filtered <- data %>% 
  filter(End.year >= 2024) %>%
  select(End.year, Award.amount, PI)
  
# Replace years after 2024 with "After 2024"
data_filtered$End.year <- ifelse(data_filtered$End.year > 2024, "After 2024", as.character(data_filtered$End.year))

# Sort by Award.amount in descending order
data_filtered <- data_filtered %>%
  arrange(desc(Award.amount))

# Convert PI to a factor variable and order levels by Award.amount
data_filtered$PI <- factor(data_filtered$PI, levels = unique(data_filtered$PI))

# Draw a column chart
ggplot(data_filtered, aes(x = PI, y = Award.amount, fill = factor(End.year))) +
  geom_col() +
  #geom_text(aes(label = Award.amount), vjust = -0.5, size = 3) +
  scale_fill_manual(values = c("2024" = "red", "After 2024"="blue"), name = "End Year") +
  labs(title = "Fig4. NIH Awards to University of Idaho (2024 and Beyond)",
       x = "Principal Investigator", y = "Award Amount") +
  theme_minimal() +
  theme(axis.text.x = element_text(angle = 45, hjust = 1),
        legend.position = "bottom",
        plot.title = element_text(hjust = 0.5, face = "bold"))

Q2

Code 
Sponsers <- c("DOE", "NIH", "NSF", "USDA")

for (Sponser in Sponsers) {

  # Read the CSV file
  data <- read.csv(paste0(Sponser, "-all.csv"), header = TRUE, stringsAsFactors = FALSE)

  # Group by Start year, calculate the total Award amount and number of projects per year
  result <- aggregate(cbind(Award.amount, n = 1) ~ University + Start.year, data = data, FUN = sum)

  # Rename columns
  colnames(result) <- c("University", "Start.year", "Award.amount", "n")

  # Sort by University and Start.year
  result <- result[order(result$University, result$Start.year), ]

  # Output results
  print(result)

  write.csv(result, file = paste0("E:/01-文件/04-UI文件/2401-【课程】BCB/BCB-Mid/00-Temp/Q2/", Sponser, "-all-Q2.csv"), row.names = FALSE)

}

# Read the CSV file
data <- read.csv("E:\\01-文件\\04-UI文件\\2401-【课程】BCB\\BCB-Mid\\00-Temp\\Q2\\UI统计表.csv", header = TRUE, stringsAsFactors = FALSE)

# Load the ggplot2 package
library(ggplot2)

# Filter data from 2013 to 2023
data_filtered <- subset(data, Start.year >= 2013 & Start.year <= 2023)

# Draw a line chart
ggplot(data_filtered, aes(x = Start.year, y = n, color = Sponser, linetype = Sponser)) +
  geom_line(size = 1) +
  scale_color_manual(values = c("USDA" = "blue", "NSF" = "green", "NIH" = "orange", "DOE" = "red")) +
  scale_linetype_manual(values = c("USDA" = "solid", "NSF" = "dashed", "NIH" = "dotted", "DOE" = "dotdash")) +
  labs(x = "Year", y = "Number of New Awards", title = "Fig5: New Awards from Different Sponsors") +
  theme_minimal() +
  theme(plot.title = element_text(hjust = 0.5, size = 16, face = "bold"),
        axis.title = element_text(size = 14),
        axis.text = element_text(size = 12),
        legend.title = element_blank(),
        legend.text = element_text(size = 12),
        legend.position = "bottom") +
  scale_x_continuous(breaks = seq(2013, 2023, by = 1))

Q3

Code 
# Define list of file names
file_names <- c("DOE-all-Q2.csv", "NIH-all-Q2.csv", "NSF-all-Q2.csv", "USDA-all-Q2.csv")

# Define list of output file names
output_names <- c("DOE-all-Q3.csv", "NIH-all-Q3.csv", "NSF-all-Q3.csv", "USDA-all-Q3.csv")

# Loop through each file
for (i in seq_along(file_names)) {
  # Read the file
  data <- read.csv(paste0("E:\\01-文件\\04-UI文件\\2401-【课程】BCB\\BCB-Mid\\00-Temp\\Q3\\", file_names[i]), header = TRUE, stringsAsFactors = FALSE)
  
  # Create a new data frame to store results
  result <- data.frame(University = character(), Start.year = numeric(), Difference.Value = numeric(), stringsAsFactors = FALSE)
  
  # Get data for University of Idaho
  ui_data <- data[data$University == "University of Idaho", ]
  
  # Calculate Difference.Value for each university
  for (j in unique(data$Start.year)) {
    year_data <- data[data$Start.year == j, ]
    ui_mean <- ui_data$Mean[ui_data$Start.year == j]
    
    for (k in unique(year_data$University)) {
      if (k != "University of Idaho") {
        university_mean <- year_data$Mean[year_data$University == k]
        difference_value <- university_mean - ui_mean
        
        result <- rbind(result, data.frame(University = k, Start.year = j, Difference.Value = difference_value))
      }
    }
  }
  
  # Write results to a CSV file
  write.csv(result, paste0("E:\\01-文件\\04-UI文件\\2401-【课程】BCB\\BCB-Mid\\00-Temp\\Q3\\", output_names[i]), row.names = FALSE)
}

# Load required packages
library(ggplot2)
library(dplyr)

# Define list of file names
file_names <- c("DOE-all-Q3.csv", "NIH-all-Q3.csv", "NSF-all-Q3.csv", "USDA-all-Q3.csv", "Overall-all-Q3.csv")

# Define list of output image names
output_names <- c("DOE_plot.png", "NIH_plot.png", "NSF_plot.png", "USDA_plot.png", "Overall-all-Q3.png")

# Define list of plot titles
plot_titles <- c("Fig6", "Fig7", "Fig8", "Fig9", "Fig10")

# Loop through each file
for (i in seq_along(file_names)) {
  # Read the file
  data <- read.csv(paste0("E:\\01-文件\\04-UI文件\\2401-【课程】BCB\\BCB-Mid\\00-Temp\\Q3\\", file_names[i]), header = TRUE, stringsAsFactors = FALSE)
  
  # Filter data for the last 5 years (2019-2023)
  data_filtered <- data %>% filter(Start.year >= 2019 & Start.year <= 2023)
  
  # Convert University column to a factor variable and specify the order
  data_filtered$University <- factor(data_filtered$University, levels = c("Boise State University", "Idaho State University", "Montana State University", "University of Montana", "Washington State University"))
  
  # Create a bar plot
  plot <- ggplot(data_filtered, aes(x = University, y = Difference.Value, fill = factor(Start.year))) +
    geom_bar(stat = "identity", position = "dodge") +
    labs(title = paste0(plot_titles[i], ": Difference in Mean Award Amount (", gsub("-all-Q3.csv", "", file_names[i]), ")"),
         x = "University",
         y = "Difference Value",
         fill = "Start Year") +
    theme_minimal() +
    theme(plot.title = element_text(hjust = 0.5, size = 16, face = "bold"),
          axis.text.x = element_text(angle = 0),
          legend.position = "bottom") +
    scale_fill_discrete(name = "Start Year")
  
  print(plot)
  # Save the image
  ggsave(paste0("E:\\01-文件\\04-UI文件\\2401-【课程】BCB\\BCB-Mid\\00-Temp\\Q3\\", output_names[i]), plot, width = 10, height = 6, dpi = 300)
}