UseR2015.Rmd

---
title: "UseR2015"
author: "Mario Santoro"
date: "06 luglio 2015"
output: ioslides_presentation
---

## Rcpp http://www.rcpp.org

Codice Performante in C++ per R

foo.cpp 
```
  #include <Rcpp.h>
  using namespace Rcpp;
  // [[Rcpp::export]]
  NumericVector timesTwo(NumericVector x) {
    return x * 2;
  }
```

## sourceCpp()

```{r,eval=TRUE}
library(Rcpp)
sourceCpp("foo.cpp")
timesTwo(2)
```

## Altre release specifiche di Rcpp
[RcppArmadillo](http://dirk.eddelbuettel.com/code/rcpp.armadillo.html)
Uso di Rcpp con la libreria di linear algebra in C++ [Armadillo](http://arma.sf.net/) 

```   
  #include <RcppArmadillo.h>
```

[RcppGSL](http://dirk.eddelbuettel.com/code/rcpp.gsl.html) interfaccia a [GNU GSL](http://www.gnu.org/software/gsl/)
```
  #include <RcppGSL.h>
  #include <gsl/gsl_matrix.h>
  #include <gsl/gsl_blas.h>
```

[RcppSMC](http://dirk.eddelbuettel.com/code/rcpp.smc.html) implementa Rcpp per  Sequential Monte Carlo and Particle Filters [(SMC)](http://www.jstatsoft.org/v30/i06)

## Rcpp [Parallel](https://github.com/RcppCore/RcppParallel)
```
  #include <RcppParallel.h>
```
```
#include <Rcpp.h>
using namespace Rcpp;

#include <algorithm>

// [[Rcpp::export]]
double vectorSum(NumericVector x) {
  return std::accumulate(x.begin(), x.end(), 0.0);
}

```

## Rcpp [Parallel](https://github.com/RcppCore/RcppParallel)
```
// [[Rcpp::depends(RcppParallel)]]
#include <RcppParallel.h>
#include <Rcpp.h>
using namespace RcppParallel;
struct Sum : public Worker
{   
  // source vector
  const RVector<double> input;
  // accumulated value
  double value;
  // constructors
  Sum(const Rcpp::NumericVector input) : input(input), value(0) {}
  Sum(const Sum& sum, Split) : input(sum.input), value(0) {}
  // accumulate just the element of the range I've been asked to
  void operator()(std::size_t begin, std::size_t end) {
    value += std::accumulate(input.begin() + begin, input.begin() + end, 0.0);
  }
  // join my value with that of another Sum
  void join(const Sum& rhs) { 
    value += rhs.value; 
  }
};
```

## Rcpp [Parallel](https://github.com/RcppCore/RcppParallel)

```
using namespace RcppParallel;
// [[Rcpp::export]]
double parallelVectorSum(Rcpp::NumericVector x) {
  // declare the SumBody instance 
  Sum sum(x);
  // call parallel_reduce to start the work
  parallelReduce(0, x.length(), sum);
  // return the computed sum
  return sum.value;
}
```

## Rcpp [Parallel](https://github.com/RcppCore/RcppParallel)
```{r,eval=TRUE}
library(RcppParallel)
sourceCpp("foo_ser.cpp")
sourceCpp("foo_par.cpp")
library(rbenchmark)
v <- as.numeric(c(1:10000000))
pp <- benchmark(vectorSum(v),parallelVectorSum(v),order="relative")
pp
```

## RInside [rinside.html](http://dirk.eddelbuettel.com/code/rinside.html)
R embedded in classi C++
```
  #include <RInside.h>
```

Pacchetto giovane, alle versione 0.2, ma promettente ...

```
#include <RInside.h>   // for the embedded R via RInside

int main(int argc, char *argv[]) {

    RInside R(argc, argv);  // create an embedded R instance

    R["txt"] = "Hello, world!\n";	// assign a char* (string) to 'txt'

    R.parseEvalQ("cat(txt)"); // eval the init string, ignoring any returns

    exit(0);
}
```


## [HTMLWIDGETS](http://www.htmlwidgets.org)
Widgets interattivi costruiti con R e javascript
Già ci sono alcuni paccheti di R che implementano librerie js

esempi con questo dataset
```{r,echo=FALSE}

library(dplyr,quietly = T,warn.conflicts = F)
library(gplots,quietly = T,warn.conflicts = F)
library(zoo,quietly = T,warn.conflicts = F)
library(xts,quietly = T,warn.conflicts = F)
library(reshape2,quietly = T,warn.conflicts = F)

# Slurp up ./data/*.csv into one data frame
downloads <- do.call(rbind, lapply(list.files("./ggbrush/data", full.names = TRUE), read.csv, stringsAsFactors = FALSE))
# Round time to nearest hour
downloads <- downloads %>% mutate(hour = sub(":.*", ":00", time)) %>% tbl_df()

# Counts per date/hour by country
counts_by_country <- downloads %>%
  count(date, hour, country) %>%
  arrange(date, hour, country)
#counts_by_country

# Counts per date/hour, all countries combined
counts <- counts_by_country %>%
  group_by(date, hour) %>%
  summarise(n = sum(n))
#counts

# Countries with at least 300 downloads
major_countries <- (downloads %>% count(country) %>% filter(n > 300))$country

# Make a matrix of countries vs. hours
m_hours_countries <- counts_by_country %>%
  filter(country %in% major_countries) %>%
  group_by(hour, country) %>%
  summarise(n = sum(n)) %>%
  arrange(hour, country) %>%
  acast(hour ~ country, value.var = 'n', fill = 0)
```
Scaricamenti orari di pacchetti R per nazione
```{r,}
m_hours_countries[1:4,1:10]
```

## [dygrhaps](http://dygraphs.com)
Libreria per serie temporali
```{r,echo=FALSE}
library(dygraphs)
# Turn into time series
times <- paste0(counts$date, " ", counts$hour, ":00 UTC") %>% as.POSIXct()
ts_times <- xts(counts$n, order.by = times, frequency = 24)
#Plot
dg <- dygraph(ts_times) %>%
  dyAxis("y", "Downloads") %>%
  dyAxis("x", "Date")
dg %<>% dyRangeSelector() %>% dyAnnotation(as.POSIXct("2015-05-17 19:00"), "A") %>%
  dyShading(as.POSIXct("2015-05-17 05:00"),
    as.POSIXct("2015-05-18 09:00")) %>%
  dyRoller()
dg
```

## [d3hetamap](https://github.com/rstudio/d3heatmap)
Heatmap usando D3
```{r,fig.height=3}
library(d3heatmap)
m_counts <- t(acast(counts, date ~ hour, value.var = 'n'))
d3heatmap(scale(m_hours_countries), cellnote = m_hours_countries,
  colors = "Blues", dendrogram = "col")
```

## [leaflet](http://leafletjs.com)
Mappe interattive
```{r,eval=F}
library(leaflet)
leaflet(quakes) %>%
  addTiles("https://api.tiles.mapbox.com/v4/mapbox.wheatpaste/{z}/{x}/{y}.png?access_token=pk.eyJ1Ijoic2FudG9yb21hIiwiYSI6ImUxN2ZkZWRkNTU2NjVlZjRiM2E2ZmZhNGQyODM0OGFhIn0.mM32XaR8Ro0dBQIlA6G2iA") %>%
  addCircles(color = "#CC0000", weight = 2, radius = ~10^mag / 5,
    popup = ~as.character(stations))
```

## [leaflet](http://leafletjs.com)
Mappe interattive
```{r,echo=F}
library(leaflet)
leaflet(quakes) %>%
  addTiles("https://api.tiles.mapbox.com/v4/mapbox.wheatpaste/{z}/{x}/{y}.png?access_token=pk.eyJ1Ijoic2FudG9yb21hIiwiYSI6ImUxN2ZkZWRkNTU2NjVlZjRiM2E2ZmZhNGQyODM0OGFhIn0.mM32XaR8Ro0dBQIlA6G2iA") %>%
  addCircles(color = "#CC0000", weight = 2, radius = ~10^mag / 5,
    popup = ~as.character(stations))
```

## [threejs globe](http://threejs.org)
Libreria 3D in javascript
```{r,}
library(threejs)
globejs(lat = quakes$lat, long = quakes$long, 
        value = 10^quakes$mag / 10^4,atmosphere = TRUE)
```

## [epiwidgets](https://github.com/sdwfrost/epiwidgets)
Tree Viewer in js D3
```{r,eval=F}
# devtools::install_github("sdwfrost/epiwidgets")
library(epiwidgets)
nwk <- "(((EELA:0.150276,CONGERA:0.213019):0.230956,(EELB:0.263487,CONGERB:0.202633):0.246917):0.094785,((CAVEFISH:0.451027,(GOLDFISH:0.340495,ZEBRAFISH:0.390163):0.220565):0.067778,((((((NSAM:0.008113,NARG:0.014065):0.052991,SPUN:0.061003,(SMIC:0.027806,SDIA:0.015298,SXAN:0.046873):0.046977):0.009822,(NAUR:0.081298,(SSPI:0.023876,STIE:0.013652):0.058179):0.091775):0.073346,(MVIO:0.012271,MBER:0.039798):0.178835):0.147992,((BFNKILLIFISH:0.317455,(ONIL:0.029217,XCAU:0.084388):0.201166):0.055908,THORNYHEAD:0.252481):0.061905):0.157214,LAMPFISH:0.717196,((SCABBARDA:0.189684,SCABBARDB:0.362015):0.282263,((VIPERFISH:0.318217,BLACKDRAGON:0.109912):0.123642,LOOSEJAW:0.397100):0.287152):0.140663):0.206729):0.222485,(COELACANTH:0.558103,((CLAWEDFROG:0.441842,SALAMANDER:0.299607):0.135307,((CHAMELEON:0.771665,((PIGEON:0.150909,CHICKEN:0.172733):0.082163,ZEBRAFINCH:0.099172):0.272338):0.014055,((BOVINE:0.167569,DOLPHIN:0.157450):0.104783,ELEPHANT:0.166557):0.367205):0.050892):0.114731):0.295021)"
treewidget(nwk)
```

## [epiwidgets](https://github.com/sdwfrost/epiwidgets)
Tree Viewer in js D3
```{r,echo=F}
# devtools::install_github("sdwfrost/epiwidgets")
library(epiwidgets,quietly = T,warn.conflicts = F)
nwk <- "(((EELA:0.150276,CONGERA:0.213019):0.230956,(EELB:0.263487,CONGERB:0.202633):0.246917):0.094785,((CAVEFISH:0.451027,(GOLDFISH:0.340495,ZEBRAFISH:0.390163):0.220565):0.067778,((((((NSAM:0.008113,NARG:0.014065):0.052991,SPUN:0.061003,(SMIC:0.027806,SDIA:0.015298,SXAN:0.046873):0.046977):0.009822,(NAUR:0.081298,(SSPI:0.023876,STIE:0.013652):0.058179):0.091775):0.073346,(MVIO:0.012271,MBER:0.039798):0.178835):0.147992,((BFNKILLIFISH:0.317455,(ONIL:0.029217,XCAU:0.084388):0.201166):0.055908,THORNYHEAD:0.252481):0.061905):0.157214,LAMPFISH:0.717196,((SCABBARDA:0.189684,SCABBARDB:0.362015):0.282263,((VIPERFISH:0.318217,BLACKDRAGON:0.109912):0.123642,LOOSEJAW:0.397100):0.287152):0.140663):0.206729):0.222485,(COELACANTH:0.558103,((CLAWEDFROG:0.441842,SALAMANDER:0.299607):0.135307,((CHAMELEON:0.771665,((PIGEON:0.150909,CHICKEN:0.172733):0.082163,ZEBRAFINCH:0.099172):0.272338):0.014055,((BOVINE:0.167569,DOLPHIN:0.157450):0.104783,ELEPHANT:0.166557):0.367205):0.050892):0.114731):0.295021)"
treewidget(nwk)
```

## Thematic Maps [tmap](https://github.com/mtennekes/tmap)
```{r}
library(tmap,quietly = T,warn.conflicts = F)
data(World)
data(metro)
qtm(World, fill="pop_est_dens", theme="World", fill.style="kmeans", 
    fill.title="Population per km^2") +
  qtm(metro, bubble.size = "pop2010", bubble.col ="purple", 
      bubble.title.size="Metropolitan Areas", theme = "World", bubble.scale=.5)
```