Introduction:

In this project, I delved into a dataset containing patient risk profiles, with an aim to visualize the predicted risk of various diseases across different age groups and medical histories. This analysis is part of my ongoing contributions to TidyTuesday, a weekly data project within the R community. The dataset for this week has been curated as part of a larger healthcare analytics initiative and is made accessible through the tidytuesday GitHub repo.

A big thank you to Jenna Reps for this data.

You can find the code from my analysis on here.

Here’s a step-by-step breakdown of my approach:

Packages:

We are going to be using tidyverse, RColorBrewer and shiny for this exercise

library(tidyverse)
── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
✔ dplyr     1.1.3     ✔ readr     2.1.4
✔ forcats   1.0.0     ✔ stringr   1.5.0
✔ ggplot2   3.4.3     ✔ tibble    3.2.1
✔ lubridate 1.9.3     ✔ tidyr     1.3.0
✔ purrr     1.0.2     
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()
ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
library(RColorBrewer)
library(shiny)

Data Loading:

The data is loaded from a URL using the read_csv function from the readr package.

# Load the data
patient_risk_profiles <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2023/2023-10-24/patient_risk_profiles.csv')
Rows: 100 Columns: 100
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
dbl (100): personId, age group:  10 -  14, age group:  15 -  19, age group: ...

ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Data Preperation:

I wanted to Reshape the data using pivot_longer to create a more analysis-friendly structure, separating out the various attributes into distinct columns and also perform necessary data cleaning and manipulation to ensure consistency and readability.

  • The df object is created by manipulating patient_risk_profiles using a series of functions from the tidyverse suite of packages.

  • The pivot_longer function is used multiple times to convert the data from a wide format to a longer format, making it easier to work with.

  • The select function is used to drop certain columns from the dataset that are not needed for further analysis.

  • The mutate function is used twice to clean and reformat the data. In the first mutate call, the Sex column is cleaned to have simpler, more readable values i.e., only have female and male represented as F and M in the dataset. In the second mutate call, unwanted text is removed from the Disease column values to make them more readable.

########## 
#### 

df <- patient_risk_profiles %>% 
  tidyr::pivot_longer(cols = starts_with("age group:"),
                      names_to = "age_group",
                      values_to = "value_age_group") %>%  
  tidyr::pivot_longer(cols = `Acetaminophen exposures in prior year`:`Antibiotics Tetracyclines in prior year`,
                      names_to = "record_of",
                      values_to = "value_record_of_group") %>% 
  tidyr::pivot_longer(cols = starts_with("predicted risk of"),
                      names_to = "Disease",
                      values_to = "Predicted_risk") %>% 
  tidyr::pivot_longer(cols = `Sex = FEMALE`:`Sex = MALE`,
                      names_to = "Sex",
                      values_to = "values_sex") %>% 
  dplyr::select(-c("value_age_group","values_sex","value_record_of_group","personId"))  %>% 
  dplyr::mutate(Sex = case_when(Sex == "Sex = FEMALE" ~ "F",
                                Sex == "Sex = MALE" ~ "M")) %>% 
  dplyr::mutate(Disease = str_remove_all(Disease, "predicted risk of "))

Shiny application creation:

UI Code Breakdown:

The ui code in Shiny defines the user interface elements of the application.

  1. UI Definition:

    • fluidPage is used as a container to hold the UI elements.

    • selectInput creates dropdown menus for user selection. Three dropdown menus are created for age group, record of medical history, and disease respectively.

    • plotOutput defines a space in the UI to render the dynamic bar plot generated in the server function.

# Define the UI
ui <- fluidPage(
  selectInput("age_group", "Age group:", unique(df$age_group)),
  selectInput("record_of", "Record of:", unique(df$record_of)),
  selectInput("Disease", "Predicted risk:", unique(df$Disease)),
  plotOutput("barPlot")
)
  1. Server function:

    • filtered_data is defined as a reactive object. This object will update automatically whenever the input values change.

    • The reactive function encapsulates code that generates reactive conduct - in this case, filtering the data based on user input.

    • output$barPlot is defined using the renderPlot function, which generates a dynamic bar plot based on the current state of filtered_data.

    • Inside renderPlot, a ggplot object is created using the filtered data, with aesthetic mappings, geoms, and other components specified to build the plot.

# Define the server
server <- function(input, output) {
  
  # Create a reactive object to filter the data based on user input
  filtered_data <- reactive({
    df %>%
      filter(age_group == input$age_group,
             record_of == input$record_of,
             Disease == input$Disease)
  })
  
  # Create the dynamic bar plot
  output$barPlot <- renderPlot({
    ggplot(data = filtered_data(), aes(x = Sex, y = Predicted_risk, fill = Sex)) +
      geom_bar(stat = "identity", position = position_dodge()) +
      scale_fill_manual(values = brewer.pal(2, "Set3")) +  # Set3 is a color-blind friendly palette
      labs(title = paste("Predicted Risk of", input$Disease,
                         "for Age Group", input$age_group,
                         "with Record of", input$record_of),
           y = "Predicted Risk",
           x = "Sex") +
      theme_minimal()
  })
  
}

Run the app:

# Run the app
shinyApp(ui = ui, server = server)

Shiny applications not supported in static R Markdown documents