https://tomcab-1988.shinyapps.io/HeartShiny/
"<3 Shiny" - why? Because the application allows for interrogation of factors relating to cardiovascular health outcomes for patients who have experienced a heart attack + because we love shiny of course.
- the data is mock data.
- built in a shinydashboard context as i've found it is the most intuitive for creating multiple tab items.
- All code written by Dr Tom Hunter who consents to its wider use, adaptation, and monetisation.
- shiny_df = ~73k or mock cardiovascular data which has undergone significant imputation and then averaging using the MICE package
- shiny_df_root = ~73k rows of the same data without imputation, to allow for interrogation of the effectiveness of the imputation process.
This dataset contains various health-related data points recorded for patients attending the hospital. Below is a description of each variable:
- There is a basic naming convention in the app such that a grouping variable starts with "Demographic_" and a numeric variable starts with "Num_"
- This is both for the user but also allows for automatic variable input as long as the modder sticks to this naming convention when feature engineering with the dataset.
Description: A unique patient identifier derived from their NHS number. This ensures anonymity while maintaining a unique reference for each patient.
Type: String or Numeric (depending on your implementation). Patient_sex:
Type: Categorical (e.g., "Male", "Female"). Arrival_at_hospital:
Description: The date the patient was admitted to the hospital. It provides insight into the timeline of the patient's care.
Type: Date (format: YYYY-MM-DD). STEMI:
Description: Refers to the type of heart attack the patient experienced. "STEMI" (ST-Elevation Myocardial Infarction) and "nSTEMI" (Non-ST Elevation Myocardial Infarction) are two different forms of heart attacks.
Type: Categorical (e.g., "STEMI", "nSTEMI"). Blood_glucose:
Description: The blood glucose level of the patient at the time of admission, measured in mmol/L. This is a critical indicator of the patient’s metabolic state.
Type: Numeric (e.g., 4.5, 7.8). Height:
Description: The patient’s height measured in centimeters (cm). This is often used in conjunction with weight to calculate BMI (Body Mass Index).
Type: Numeric (e.g., 172, 185). Weight:
Description: The patient’s weight in kilograms (kg), which can be used alongside height to assess health metrics such as BMI.
Type: Numeric (e.g., 70.5, 85.3). Creatinine:
Description: The patient’s blood creatinine level at admission, measured in micromoles per liter (umol/L). Elevated levels could indicate kidney dysfunction.
Type: Numeric (e.g., 75, 110). Cholesterol:
Description: The patient’s total serum cholesterol level, measured in mmol/L. High levels are associated with an increased risk of cardiovascular diseases.
Type: Numeric (e.g., 4.5, 6.2). Systolic_bp:
Description: The patient’s systolic blood pressure at admission, measured in millimeters of mercury (mmHg). This represents the pressure in the arteries during heart contractions.
Type: Numeric (e.g., 120, 140). DTB (Door-to-Balloon Time):
Description: The time in minutes from the patient’s arrival at the hospital (door) to the inflation of a balloon in the coronary artery (balloon), in cases of angioplasty. This is relevant only for patients who received this procedure.
Type: Numeric (e.g., 90, 120). Death_in_hospital:
Description: A binary variable indicating whether the patient died in the hospital. A value of 1 means the patient died in the hospital, and 0 indicates survival.
Type: Binary (1 for death, 0 for survival). Required Libraries The following R libraries are used in the application to handle data, create visualizations, and manage the user interface:
library(shinydashboard) # For creating a dashboard layout in Shiny. library(shiny) # The main Shiny library for building the app. library(DT) # For rendering interactive tables. library(tidyverse) # A collection of packages for data manipulation and visualization. library(kableExtra) # For generating tables in HTML and LaTeX. library(reshape2) # For reshaping data (e.g., melting and casting). library(janitor) # For cleaning data and creating tidy data frames. library(zoo) # For working with ordered observations (especially time series). library(rlang) # Tools for advanced programming with R. library(glue) # For interpolating strings. library(stringr) # For string manipulation. library(plotly) # For creating interactive plots. library(fresh) # For adding custom themes to the Shiny dashboard. library(sass) # For writing and compiling SCSS. library(sf) # For working with geospatial data. library(shinyWidgets) # For enhancing the Shiny interface with additional widgets. library(writexl) # For writing data to Excel files. library(logger) # For logging messages during the app runtime. Menu Items (Tabs) The application has multiple tabs, each representing a different data view or functionality. Below is a list of the menu items and their corresponding tab names: Front Page: The home tab of the application.
The application consists of several tabs, each providing different functionalities:
Front Page: The home tab of the application. Frequencies: Displays frequency counts of selected variables. Proportions: 1 Group: Shows proportions for a single group. Proportions: 2 Groups: Displays comparative proportions for two groups. Averages: 1 Group: Presents average values for a single group. Averages: 2 Groups: Shows comparative average values for two groups. Line: 1 group: Line plot for a single group over time. Line: 2 groups: Comparative line plot for two groups over time. Bar: Simple: Simple bar chart visualization. Bar: Stacked: Stacked bar chart to show proportions within categories. Bar: Grouped: Grouped bar chart for side-by-side comparison. Area: Stacked: Stacked area chart visualization. Area: %: Displays area chart as percentage proportions. Scatterplots: Scatter plot for grouped data. Density: Density plot for visualizing data distributions.
For running the application from your local Windows machine, you can use the following method via a .bat file. Here’s a simplified workflow for doing so:
Ensure that you have R installed on your machine. Use a Windows batch file to start the application locally. A sample batch file to run the app:
swift Copy code @echo off set R_SCRIPT="C:\path\to\R\bin\x64\R.exe" set APP_DIR=C:\path\to\your\app\directory
%R_SCRIPT% -e "options(error = function() {cat('An error has occurred:\n', geterrmessage(), '\n'); q('no')});" %R_SCRIPT% -e "shiny::runApp('%APP_DIR%', launch.browser = TRUE)" 2>&1 pause