This repository contains all of the files required for an analysis of the Toronto Paramedic Services' response times. The aim of this project is to analyse the Toronto Paramedic Services' response times to determine its efficacy as an emergency medical service.
This project requires both the R programming language and Quarto. If you do not have these tools in your development environment, please install them now. You will also need an integrated development environment (IDE) capable of running R scripts. I recommend RStudio (local) or Posit Cloud (cloud-based).
Once your environment is set up, you must install several packages that handle various tasks, like graphing data, creating tables, and general organization and processing. You will find a complete list of these packages in the file scripts/00-install_dependencies.qmd. On lines 16-29 of 00-install_dependencies.qmd, there is a code chunk that installs all of the required dependencies for this project. You only need to run this code once to install the required dependencies.
The first step in working with this project is to download three data sets from Open Data Toronto. Don't worry, there's no need for you to physically download any files; the code will take care of that.
The scripts/01-data_import.qmd file will download the required data sets for you. This file is a Quarto document, which is a technical publishing system using Markdown as a foundation and allowing us to run embedded R code chunks. Within 01-data_import.qmd, you'll see three separate R code chunks:
- The first chunk (lines 20-34) downloads the Land Ambulance Response Time Standard package
- The second chunk (lines 36-51) downloads the Pre-Hospital Emergency Care Performance Metrics package
- The third chunk (lines 53-69) downloads the Paramedic Services Incident Data package
Run each chunk to fetch the data sets and import the relevant data into .csv files. At the end of this process, you should have three new files in inputs/data: raw_response_data.csv, raw_incident_data.csv, and raw_performance_data.csv.
Before moving to data analysis, we must clean the generated .csv files to help us filter, use, and understand the relevant data points. The scripts/02-data_cleaning.qmd file handles all of the data cleaning, including fixing column names (many have characters that cannot be used or are insufficent descriptors), selecting the appropriate columns, and filtering any rows that contain null data. Within 02-data_cleaning.qmd, you'll see three separate R code chunks:
- The first chunk (lines 21-28) cleans
inputs/data/raw_response_data.csv - The second chunk (lines 40-57) cleans
inputs/data/raw_performance_data.csv - The third chunk (lines 59-73) cleans
inputs/data/raw_incident_data.csv
Run each chunk to fetch the raw data sets, clean them, and then create new .csv files with the clean data. At the end of this process, you should have three new files in inputs/data: clean_response_data.csv, clean_incident_data.csv, and clean_performance_data.csv.
The core data analysis of this project occurs in the outputs/paper/paper.qmd file, another Quarto document. Once you render paper.qmd, Quarto will generate a paper.pdf file in the same folder. The raw references used in paper.qmd are available under the same folder in the references.bib file.
If you're experiencing problems with the data, I've compiled a document that tests the data against several parameters, like data types, number ranges, and data ranges. This testing document is available under the scripts/03-data_testing.qmd file. The file contains three code chunks testing each imported package from Open Data Toronto:
- The first chunk (lines 22-38) tests data from
inputs/data/raw_response_data.csv - The second chunk (lines 40-56) tests data from
inputs/data/raw_performance_data.csv - The third chunk (lines 58-78) tests data from
inputs/data/raw_incident_data.csv
Before running any of these tests, you must first download the data following the steps outlined above. All of these tests should return true. If they do not, feel free to create an issue.
If you'd like to debug the problem yourself, or if you'd like to use a service like Stack Overflow for help, it's important to have some simulated data to reproduce the problem. The scripts/04-data_simulation.qmd file generates random, fake data based on the information initially downloaded from Open Data Toronto.
I opted to use Quarto files instead of R scripts to handle processes like data testing and simulation. This allowed me to divide tasks into separate chunks capable of running independently. Although this is not standard convention, I felt it would make this project more reproducible by allowing users to identify specific errors in each chunk. By allowing code chunks to run independently, it also gives users the opportunity to continue with a process despite errors (for example, if users encounter an error with the code in one test, it has no impact on the other tests).
Created by Sebastian Rodriguez © 2023, licensed under the BSD 3-Clause License. Contains information from Open Data Toronto, licensed under the Open Government Licence - Toronto. Created using R, an open-source statistical programming language.
This project uses a number of R packages, including: dplyr, ggplot2, here, janitor, kableExtra, knitr, lubridate, opendatatoronto, readr, RColorBrewer, scales, and tidyverse.
Much of this project's development was informed by Rohan Alexander's book Telling Stories with Data.