Next steps

class: center, middle, inverse, title-slide

# Next steps
### The R Bootcamp <a href='https://therbootcamp.github.io'>www.therbootcamp.com</a> <a href='https://twitter.com/therbootcamp'>@therbootcamp</a>
### July 2018

---

# Next steps

.pull-left4[
What we would have liked to cover 

1. Networks 
2. Bayesian stats 
3. Text analysis 
4. Statistics 
5. Rcpp 
6. Forms 
7. Github 
8. R on servers

How to continue 

8. Books 
9. Websites R 
10. Groups 

]

.pull-right5[
<img src="https://i0.wp.com/www.pieceofk.fr/wp-content/uploads/2017/09/Rplot.png?zoom=2&w=450" width=400 align="center"> 
source <a href="https://www.r-bloggers.com/exploring-the-cran-social-network/">R-bloggers.com</a>

]

---

# Network analysis

A <high>social graph</high> of package Co-authors using `tidyverse` plus `ggraph`, an extension for `ggplot2` for graphs (aka networks) and `igraph`, an extremely powerful network analysis library. Find the code and additional explanation [here](https://www.r-bloggers.com/exploring-the-cran-social-network/).

.pull-left[
 <img src="https://i1.wp.com/www.pieceofk.fr/wp-content/uploads/2017/09/Rplot04.png?zoom=2&w=450" width=450> 
 source <a href="https://www.r-bloggers.com/exploring-the-cran-social-network/">R-bloggers.com</a>
 ]

.pull-right[

 
 <img src="https://i1.wp.com/www.pieceofk.fr/wp-content/uploads/2017/09/Rplot02.png?zoom=2&w=450" width=450> 
 source <a href="https://www.r-bloggers.com/exploring-the-cran-social-network/">R-bloggers.com</a>
 ]

---

# Semantic networks - [www.swow.me](https://www.smallworldofwords.org/de)

---

# Text analysis

Another area were R has seen much development is <high>text analysis</high>. New packages provide access to lightning fast <high>C++ string libraries</high>, 'new' data formats such as <high>`html`</high>, <high>xml</high>, and <high>json</high>, and also to the API of social platforms, such as [facebook.com](www.facebook.com) and [twitter.com](www.twitter.com).

| Package| Description| 
|:------|:------------|
| `tm`, `tidytext`  | General text analysis packages  |
| `stringr`, `stringi`  | String operations and regular expressions. |
| `rvest`, `XML`  | Scraping content of the internet|
| `text2vec`| Create vector representation of words. |
| `SentimentAnalysis`| Sentiment analysis.|
| `twitteR`, `streamR`, `jsonlite`| Streaming and parsing tweets. |
| `Rfacebook`| Accessing the Facebook API. |

---

# Text analysis: Applications - [hedonometer.org](http://hedonometer.org/books/v1/)
<br2>
<a href="http://hedonometer.org/books/v1/">
 <img src="https://raw.githubusercontent.com/therbootcamp/therbootcamp.github.io/master/_sessions/_image/harry_potter.png" width=900></a> 
 Link to <a href="https://arxiv.org/pdf/1606.07772.pdf">Paper</a> 
 View <a href="https://www.youtube.com/watch?v=oP3c1h8v2ZQ">Vonnegut's Shape of Stories</a>

---

# Stats packages

>"If there is a new statistical method, it is first implemented in R" 
> -- duw

| Package| Description| 
|:------|:------------|
| `stats`  | Many individual tests plus all the distributions in the world.|
| `lme`, `glm`  | (Generalized) linear models.|
| `lme4`, `afex`  | Mixed-mode, hierarchical regression.|
| `caret`, `mlr`, `e1071`, `rpart`, etc.| Machine learning.|
| `BayesFactor`, `rstanarm`| Bayesian linear models.|
| `rjags`, `rstan`| Bayesian graphical models.|
| `forecast`, `mgm`, `timeSeries`, etc.| Time series models.|
| `cluster`, `fastcluster`, `cstab`, etc.| Cluster analysis.|

---

# Deep learning

---

# Deep learning: Tensorflow

.pull-left5[

```r
# Get Keras
install.packages(keras)
library(keras)
install_keras()

# Define model
model <- keras_model_sequential() 
model %>% 
 layer_dense(units = 256,
 activation = 'relu', 
 input_shape = c(784)) %>% 
 layer_dropout(rate = 0.4) %>% 
 layer_dense(units = 128, 
 activation = 'relu') %>%
 layer_dropout(rate = 0.3) %>%
 layer_dense(units = 10, 
 activation = 'softmax')
```
]

.pull-right45[
<a href="https://tensorflow.rstudio.com/"><img src="https://upload.wikimedia.org/wikipedia/commons/thumb/1/11/TensorFlowLogo.svg/2000px-TensorFlowLogo.svg.png" height = 400></a>
]

---

# Bayesian statistics

.pull-left5[
Recent years several convenient and efficient packages for <high>Bayesian statistics</high> have been developed, facilitating a shift to a more informative and consistent statistical framework. <br2>
Using these packages is oftentimes <high>as easy as using their Frequentist counterparts</high>. See in particular the [BayesFactor](http://bayesfactorpcl.r-forge.r-project.org/) package.

<img src="http://www.labtimes.org/labtimes/method/methods/img/2011_04b.jpg" width=350> 
 source <a href="http://www.labtimes.org/labtimes/method/methods/img/2011_04b.jpg/">labtimes.org</a>
 ]

.pull-right45[

```r
# load package
library(BayesFactor)

# define data
x <- rnorm(100, 0, 1)
y <- rnorm(100, 1, 1)

# run Frequentist analysis
t.test(x, y)

# run Bayesian analysis
ttestBF(x, y)
```

```
## Bayes factor analysis
## --------------
## [1] Alt., r=0.707 : 253308 ±0%
## 
## Against denominator:
##   Null, mu1-mu2 = 0 
## ---
## Bayes factor type: BFindepSample, JZS
```
]

---

# Rcpp

By now one of the most referenced R packages is Rcpp - <high>R's interface to C++</high>. With often relatively little effort due to <high>Rcpp sugar</high>, Rcpp can provide vast speed improvements, which many packages today rely on Rcpp in the background for <high>swift code execution</high>. Rcpp becomes particularly powerful, when supplemented with `BH`, which makes avaialble a collection <high>free, peer-reviewed C++ libraries</high>, and [RcppArmdillo](http://arma.sourceforge.net/), which available the high-performance <high>Armadillo</high> library for linear algebra methods.

.pull-left4[

 <img src="https://classic105.s3.amazonaws.com/wp-content/uploads/2017/05/180207142-sugar.jpg" width=350> 
 source <a href="https://classic105.s3.amazonaws.com/wp-content/uploads/2017/05/180207142-sugar.jpg/">classic105.com</a>
]

.pull-right6[

<img src="https://raw.githubusercontent.com/therbootcamp/therbootcamp.github.io/master/_sessions/_image/Rcpp.png" width=480> <br2>
 <a href="http://dirk.eddelbuettel.com/code/rcpp/Rcpp-quickref.pdf">Quick Reference Guide</a>
]

---

.pull-left5[
# Google Forms & Maps

New packages also allow you to interact with [Google Maps](https://www.google.com/maps) and [Google Forms](https://docs.google.com/forms). Use `ggmap` to access Google Maps and `googlesheets` to access Google Forms.

```r
library(ggmap)
ggmap(get_map(c(7.588576, 47.559601),zoom=16))
```

 <img src="https://raw.githubusercontent.com/therbootcamp/therbootcamp.github.io/master/_sessions/_image/basel.png" width=350>
]

.pull-right45[

]

---

# Github & Dropbox

An important part of programming is to <high>keep track of your project's progress</high> to potentially be able to go back to an earlier version. To achieve this, many programmers rely on <high>version-control</high> software such as [<high>GitHub</high>](www.github.com). R and, in particular, RStudio work have inbuilt software that allow you to <high>sync your project</high> with an <high>GitHub repository</high>. Via the package called `rdrop2`, R also interfaces with the popular cloud service [<high>Dropbox</high>](www.dropbox.com), which also allows for rudimentary version control.

.pull-left3[

]

.pull-right7[

]

---

# R on servers

With the need for more computing power for, e.g., machine learning, R is also moving more and more to the <high>server and cluster sphere</high>. Rstudio offers a convenient (and free) solution to access [RStudio](https://www.rstudio.com/products/rstudio/) sessions on a <high>remote linux server</high> via a browser. RStdudio also offers the `sparklyr` package integrating the [Spark](https://spark.apache.org/) architecture for efficient cluster computing with `dplyr`. Other packages exist to interface with [Hadoop](http://hadoop.apache.org/) (`RHadoop`) and [SQL](https://de.wikipedia.org/wiki/SQL) (`RSQLite`).

.pull-left6[

]

.pull-right4[

]

---

# How to continue

.pull-left5[


1. Practice 
2. Read books 
3. Consult websites 
4. Attend R group 
5. Get consulting 
]

.pull-right5[
<img border="0" alt="W3Schools" src="http://www.theunmanageableemployee.com/wp-content/uploads/2011/07/5cs-stones-cropped-11052730.jpg?w=112">
]

---

# Books

Here is a very incomplete series of good books. They are ordered by complexity, beginning with user-friendly books on <high>learning statistics</high> in R and ending with books focusing on the more <high>advanced topics of the R language</high>.

<tr>

</tr>

</table>

---

.pull-left4[

# Websites

The web is a great place to learn about R.

Best begin your journey with:

[Google](www.google.com) or [Rseek](www.rseek.org), which is a wrapper around google to maximize hits related to R. However, most of the time Google works just fine. Just be sure to add *r* to the the search query.

Then you will most likely be redirected to one of two pages

[R-bloggers](www.r-bloggers.com) is a website on which R users inform each other on the newest developments. See, e.g., Nathaniel's [entry](https://www.r-bloggers.com/the-new-and-improved-pirateplot-now-with-themes/).

[Stackoverflow](www.r-bloggers.com) is a website on which R users exchange problems and solutions to problems. Try post something yourself. You will be amazed by the turnaround.

]

.pull-right5[
 
 <a href="https://www.google.com/"><img border="0" alt="W3Schools" src="https://raw.githubusercontent.com/therbootcamp/therbootcamp.github.io/master/_sessions/_image/google.png" height="100"></a><img border="0" alt="W3Schools" src="https://raw.githubusercontent.com/therbootcamp/therbootcamp.github.io/master/_sessions/_image/comma.png" height="50"><a href="http://rseek.org/"><img border="0" alt="W3Schools" src="https://raw.githubusercontent.com/therbootcamp/therbootcamp.github.io/master/_sessions/_image/rseek.png" height="100" style="margin-left:10px; margin-buttom:10px"></a> 
 <a href="https://www.r-bloggers.com/"><img border="0" alt="W3Schools" src="https://raw.githubusercontent.com/therbootcamp/therbootcamp.github.io/master/_sessions/_image/rbloggers.png" height="105" style="margin-bottom:10px"></a> 
 <a href="https://stackoverflow.com/"><img border="0" alt="W3Schools" src="https://raw.githubusercontent.com/therbootcamp/therbootcamp.github.io/master/_sessions/_image/stackoverflow.png" height="105"></a>

]

---

# R meetup groups

Great place to meet other <high>R enthusiast</high> in and outside of industry and see interesting presentions on R's latest developments.

.pull-left5[

 <a href="https://www.meetup.com/de-DE/BaselR/"><img border="0" alt="W3Schools" src="https://raw.githubusercontent.com/therbootcamp/therbootcamp.github.io/master/_sessions/_image/baselr.png" height="400"></a>
]

.pull-right5[

]

---

# Consulting

Feel free to contact us.

Dr. Dirk Wulff 
<a href="www.dirkwulff.org">dirkwulff.org</a> 
<a href="https://github.com/dwulff">github.com/dwulff</a> 
packages: 
<a href="https://cran.r-project.org/web/packages/cstab/index.html">cstab</a>, 
<a href="https://cran.r-project.org/web/packages/mousetrap/index.html">mousetrap</a>, <a href="https://github.com/dwulff/choicepp">choicepp
 
 
 </th>
 <th class="tg-yw4l"> 
 <a href="www.nathanieldphillips.com/"><img border="0" alt="W3Schools" src="https://avatars3.githubusercontent.com/u/8480980?v=4&s=460" height="230"></a>

Dr. Nathaniel Phillips 
<a href="www.nathanieldphillips.com/">nathanieldphillips.com</a> 
<a href="https://github.com/ndphillips">github.com/ndphillips</a> 
packages: 
<a href="https://github.com/ndphillips/yarrr">yarrr</a>, 
<a href="https://github.com/ndphillips/FFTrees">FFTrees</a></th>

Markus Steiner 
<a href="https://github.com/mdsteiner">github.com/mdsteiner</a> 
package: 
<a href="https://github.com/ndphillips/ShinyPsych">ShinyPsych 
</th></tr>
</table>

---

# New courses

.pull-left[

## Introduction to Data Science with R

<high>Early October</high>, probably October 6 & 7

<high>2 days</high> (covering roughly the first 2.5 days of the bootcamp)

]

.pull-right[

## Applied Machine Learning with R

<high>Mid January</high>, probably January 11-13

<high>2-3 days</high> (covering ~ the first 2.5 days of the bootcamp)

]

---

# Thank you

<a href="http://i.imgur.com/PgtvQar.png">Here is an R Joke.</a> 
And now one more thing...