binb/inst/iqssDemo/iqssDemo.Rmd

---
author: Ista Zahn and Gary King
classoption: compress
fontsize: 12pt
title: IQSS Beamer Class Demonstration
date: \today
institute: IQSS 
output: binb::iqss
vignette: >
  %\VignetteIndexEntry{binb IQSS Demo}
  %\VignetteKeywords{binb,vignette}
  %\VignettePackage{binb}
  %\VignetteEngine{knitr::rmarkdown}
---


# Beamer Features

## Some of Gary's Examples

### What's this course about? 


::: incremental 

- \alert{Specific statistical methods for many research problems}
      - How to learn (or create) new methods
      - Inference: \underline{Using facts you know to learn about
        facts you don't know}
- \alert{How to write a publishable scholarly paper}
- \alert{All the practical tools of research} --- theory,
    applications, simulation, programming, word processing, plumbing,
    whatever is useful
- $\leadsto$ \alert{Outline and class materials:}
    - \mbox{{\huge\parbox[b][.5in][t]{1in}{\alert{j.mp/G2001}}}
        $\qquad\qquad$\includegraphics[width=.95in]{figs/phbAr.png}}
    - The syllabus gives topics, not a weekly plan.
    - We will go as fast as possible subject to everyone following along
    - We cover different amounts of material each week

:::

### How much math will you scare us with?

- All math requires two parts: \alertb{proof} and \alertb{concepts \& intuition}
- Different classes emphasize:
    - \alert{Baby Stats}: dumbed down proofs, vague intuition
    - \alert{Math Stats}: rigorous mathematical proofs
    - \alert{\underline{Practical Stats}}: deep concepts and intuition, proofs when needed
        - Goal: how to do empirical research, in depth
        - Use rigorous statistical theory --- when needed
        - Insure we understand the intuition --- always
        - Always traverse from theoretical foundations to practical applications
        - Includes ``how to'' computation
        - $\leadsto$ Fewer proofs, more concepts, better practical knowledge
- Do you have the background for this class? 

. . .

\alert{A Test: What's this?
    \begin{align*}
      b=(X'X)^{-1}X'y
    \end{align*} }


### Systematic Components: Examples

\includegraphics[width=8cm]{figs/functionalForms}

- \alertb{$E(Y_i) \equiv \mu_i = X_i\beta = \beta_0 +
      \beta_1X_{1i} +\dots+\beta_kX_{ki}$}
- \alertc{$\Pr(Y_i=1) \equiv \pi_i =
      \frac{1}{1+e^{-x_i\beta}}$}
- \alertd{$V(Y_i)\equiv \sigma_i^2 = e^{x_i\beta}$}
- Interpretation:
    - Each is a \alert{class of functional forms}
    - Set $\beta$ and it picks out one \alert{member of the class}
    - \alert{$\beta$} in each is an ``effect parameter'' vector,
      with different meaning


### Negative Binomial Derivation

\uncover<+->{Recall:}

\begin{equation*}
    \uncover<+->{\Pr(A|B)=\frac{\Pr(AB)}{\Pr(B)} \implies 
        \alertb{\Pr(AB)}=\alerte{\Pr(A|B)}\alertd{\Pr(B)}}
\end{equation*}

\alertb<1-1>{one}
\alertc<2-2>{two}
\alertd<3-3>{three}

\begin{align*}
  \uncover<+->{\text{NegBin}(y|\phi,\sigma^2) &= \int_0^\infty
    \alerte{\text{Poisson}(y|\lambda)}
    \times\alertd{\text{gamma}(\lambda|\phi,\sigma^2)}d\lambda\\}
  \uncover<+->{&= \int_0^\infty
    \alertb{\P(y,\lambda|\phi,\sigma^2) }d\lambda\\}
  \uncover<+->{&=
    \frac{\Gamma\left(\frac{\phi}{\sigma^2-1}+y_i\right)}
    {y_i!\Gamma\left(\frac{\phi}{\sigma^2-1}\right)}
    \left(\frac{\sigma^2-1}{\sigma^2}\right)^{y_i}
    \left(\sigma^2\right)^{\frac{-\phi}{\sigma^2-1}}}
\end{align*}


# Other Features

## Structural Features

### Structural Features

#### Levels of Structure

- usual \LaTeX\ \textbackslash\ section, \textbackslash\ subsection commands
- `frame` environments provide slides
- `block` environments divide slides into logical sections
- `columns` environments divide slides vertically (example later)
- overlays (\`a la prosper) change content of slides dynamically
  
#### \alertc{Overlay Alerts}

On the first overlay, \alert<1>{this text} is highlighted (or \emph{alerted}).  
On the second, \alert<2>{this text} is. 


### Code blocks

\footnotesize

```r
# Say hello in R
hello <- function(name) paste("hello", name)
```
 
. . .

```python
# Say hello in Python
def hello(name):
return("Hello" + " " + name)
```

. . .

```haskell
-- Say hello in Haskell
hello name = "Hello" ++ " " ++ name
```

. . .

```c
/* Say hello in C */
#include <stdio.h>
int main()
{
  char name[256];
  fgets(name, sizeof(name), stdin);
  printf("Hello %s", name);
  return(0);
}
```

\normalsize


### Alerts

- First level \alert{alert}
- Second level \alertb{alert}
- Third level \alertc{alert}
- Fourth level \alertd{alert}
- Fifth level \alerte{alert}

# More Features 

## Blocks

### Other Features

#### Levels of Structure

- Clean, extensively customizable visual style
- Hyperlinks ([http://github.com/izahn/iqss-beamer-theme](click here_) 
- No weird scaling prosper
    - slides are 96~mm~$\times$~128~mm
    - text is 10-12pt on slide
    - slide itself magnified with Adobe Reader/xpdf/gv to fill screen
- pgf graphics framework easy to use
- include external JPEG/PNG/PDF figures
- output directly to pdf: no PostScript hurdles
- detailed User's Manual (with good presentation advice, too)


### Theorems and Proofs
\framesubtitle{The proof uses \textit{reductio ad absurdum}.}

#### Theorem
There is no largest prime number.

#### Proof
> - Suppose $p$ were the largest prime number.
> - Let $q$ be the product of the first $p$ numbers.
> - Then $q+1$ is not divisible by any of them.
> - But $q + 1$ is greater than $1$, thus divisible by some prime
    number not in the first $p$ numbers. \qedhere

### Blocks

#### Normal block
A \alert{set} consists of elements.


#### \alert{Alert block}
$2=2$.

#### \alertc{Example block}
The set $\{1,2,3,5\}$ has four elements.


# Appendix

---

Backup Slides

<!-- \subsection{More stuff} -->

---

Details

---

Text omitted in main talk.

<!-- \subsection{Even more additional material} -->

---

More details

---

Even more details