In laderast/DSI_explore1: Exploratory Data Analysis and Statistics for OHSU Data Science Institute

%% smart %% to=latex %% template=tex/poster_eco.tex %% filter=tex/poster-filters.py %% biblatex

library('ggplot2')
library('knitr')
theme_set(theme_grey(base_size=10))
knitr::opts_chunk$set(fig.width = 16, fig.height = 4)

[columns=2]

[column]

Introduction

Motivation

• Students find statistical concepts hard to learn
• Sometimes it's hard to see the math for the trees
• Visual Explanations are useful for diverse learners

Interactive Visualizations are Even Better

• Understand relationships in data
• Static plots: 2 way, but interactive 3 way or more!

Not a Substitute for Teaching!

Lectures are combination of didactic and active learning

• Demonstrate the Concept (Didactic)
• Ask a Question ()
• Invite students to interact with the data
• Discuss observations
• Reach consensus about questions

One Example

• Simpson's Paradox - group of older women with outcomes

Ask A Question: Why are you more likely to live if you have smoked? Interaction: examine the relationship between age/death and age/smoking Discuss Observations: Age is confounded with smoking status, and death is high Reach Consensus: Relationship holds for those under 59

Methods

• LearnR - R Package for developing interactive tutorials
• Shiny - R package for developing interactive visualizations

Pros/Cons to Interactive Visualization

Pros

• Less math
• More accessible to beginners
• Makes mathematical concepts more memorable

Cons

• Can turn off more advanced students
• Requires programming skills
• Visualizations must be tested for effectiveness

Results

• Students were empowered to learn more
• Students liked the visualizations
• Students felt more engaged with the subject

Graphs

Using bookdown or rmarkdown, users can insert figures into a document in a friendly way. Just embed the R scripts into your .Rmd file(s) and compile it/them with knitr package, and the figures will appear automatically with high resolution and reproductivity.

ggplot(iris, aes(Sepal.Length, Petal.Length)) +
    geom_point(aes(color = Species)) +
    facet_wrap(~ Species)

Table

Similarly to graphs, tables can be inserted either with R script or with markdown syntax.

knitr::kable(
  head(iris, 3), caption = 'IRIS data as an example.',
  booktabs = TRUE
)

Bibliography

\printbibliography

[column]

Book types supported by bookdownplus

Some sample pages are displayed as follows. More details can be found in the tutorial book [@R-bdp].

\begin{figure}[htp] \includegraphics[width=.33\textwidth]{images/mail.pdf}\hfill \includegraphics[width=.33\textwidth]{images/mail.pdf}\hfill \includegraphics[width=.33\textwidth]{images/mail.pdf} \end{figure}

\begin{figure}[htp] \includegraphics[width=.33\textwidth]{images/mail.pdf}\hfill \includegraphics[width=.33\textwidth]{images/mail.pdf}\hfill \includegraphics[width=.33\textwidth]{images/mail.pdf} \end{figure}

\begin{figure}[htp] \includegraphics[width=.33\textwidth]{images/mail.pdf}\hfill \includegraphics[width=.33\textwidth]{images/mail.pdf}\hfill \includegraphics[width=.33\textwidth]{images/mail.pdf} \end{figure}

Take-home message

• R bookdownplus is an extension of bookdown for academic and literal writing, especially for reproducible reports.

• R bookdownplus is still being developed. Feel free to join me either in contributing templates to my Github repo ^[https://github.com/pzhaonet/bookdownplus], or in writing the tutorial of R bookdownplus [@R-bdp].

Acknowledgements

Many thanks to Dr. Yihui Xie for his amazing generosity and incredible patience in helping me get into the wonderful world of R bookdown.

The open source codes from Github and R community are much appreciated.

\vskip1.5cm

[/columns]

laderast/DSI_explore1 documentation built on May 29, 2019, 3:40 a.m.