Data Viz

Files used for the November 21, 2024 presentation the the Johns Hopkins Maternal Health Data Innovation and Coordinating Hub of the NIH Maternal Health Research Centers of Excellence.

Talk Recording

A recording of this presentation was published on YouTube. Watch Here.

Contents

The below files are included in this repository:

mhviz/
├── data/                              # Data used in the project
├── images/                            # Images used in the project
├── causal-model.Rmd                   # Generates the causal model DAGs
├── colors.R                           # Color definitions from TailwindCSS
├── correlations-visualized.Rmd        # Generates the correlation visualizations
├── data-simulations.Rmd               # Generates the simulated data
├── design.Rmd                         # Generates the design illustrations
├── misleading-viz.Rmd                 # Generates the misleading visualizations
├── sim-viz.Rmd                        # Generates visualizations of the simulated data
└── utils.R                            # Utility functions

These files are used to generate the visualizations and data used in the project. The data is simulated and should not be used for external analyses.

Data

State population data came from the 2020 US Census. They are modified by hand for our purposes and should not be used for external analyses.

Software Used

Statistical Software

• R
• RStudio
• Tidyverse

All examples are made in R using tidyverse tools. You can render the .Rmd files in RStudio.

Packages

The following R packages are used in this project. Copy below to install them.

install.packages(c(
  "broom",
  "broom.mixed",
  "causaldata",
  "cowplot",
  "dagitty",
  "dplyr",
  "forcats",
  "ggdag",
  "ggplot2",
  "ggrepel",
  "ggtext",
  "janitor",
  "kableExtra",
  "lme4",
  "maps",
  "purrr",
  "readr",
  "sf",
  "showtext",
  "stringr",
  "tidyr",
  "usmap"
))

References

• Massive credit goes to Kieran Healy for his Data Visualization: A Practical Introduction book. The code in this repository is heavily influenced by his work.
• The data correlation pattern code is adapted from Jan Vanhove.