---
title: "Pronósticos"
output: 
  flexdashboard::flex_dashboard:
    orientation: rows
    vertical_layout: scroll
    theme: lumen
    social: [ "twitter", "facebook", "menu"]
    source_code: embed
---

```{r setup, include=FALSE}
library(flexdashboard)
library(tidyverse)
library(xml2)
library(rvest)
library(purrr)
library(prophet)
library(lubridate)
library(hablar)
```

```{r}
fuente1 <- "https://coinmarketcap.com/currencies/bitcoin/historical-data/?start=20141201&end=20201017"
fuente2 <- "https://coinmarketcap.com/currencies/ethereum/historical-data/?start=20130428&end=20201017"
fuente3 <- "https://coinmarketcap.com/currencies/xrp/historical-data/?start=20130428&end=20201017"
fuente4 <- "https://coinmarketcap.com/currencies/litecoin/historical-data/?start=20130428&end=20201017"
fuente5 <- "https://coinmarketcap.com/currencies/bitcoin-cash/historical-data/?start=20130428&end=20201017"
fuente6 <- "https://coinmarketcap.com/currencies/basic-attention-token/historical-data/?start=20130428&end=20201017"
fuente7 <- "https://coinmarketcap.com/currencies/decentraland/historical-data/?start=20130428&end=20201017"
fuente8 <- "https://coinmarketcap.com/currencies/golem-network-tokens/historical-data/?start=20130428&end=20201017"
```





 



Bitcoin
=====================================

Row
-------------------------------------

### Pronóstico BTC

```{r}
datos1 <- read_html(fuente1) 
tabla1 <- datos1 %>%
  html_nodes("table") %>%
  html_table(header=T)
tabla12 <- tabla1[[3]]
tabla12$Date <- mdy(tabla12$Date)
tabla12 <- tabla12[with(tabla12, order(tabla12$Date)), ] 
colnames(tabla12) <- c("Date","Open","High","Low","Close","Volume","Market_Cap")
tabla12$Close<-as.numeric(gsub(",", "", tabla12$Close))
ds1 <- tabla12$Date
y1 <- log(tabla12$Close)
df1 <- data.frame(ds1, y1)
colnames(df1) <- c("ds","y")
m1 <- prophet(df1)
future1 <- make_future_dataframe(m1, periods = 365)
forecast1 <- predict(m1, future1)

plot(m1, forecast1)

```

> Fuente: Elaboración propia

### Tendencia BTC

```{r}
prophet_plot_components(m1, forecast1)
```

> Fuente: Elaboración propia

Row
-------------------------------

### Gráfico Interactivo BTC
```{r}
dyplot.prophet(m1, forecast1)
```

> Fuente: Elaboración propia

Ethereum
=====================================

Row
-------------------------------------

### Pronóstico ETH

```{r}
datos2 <- read_html(fuente2) 
tabla2 <- datos2 %>%
  html_nodes("table") %>%
  html_table(header=T)
tabla22 <- tabla2[[3]]
tabla22$Date <- mdy(tabla22$Date)
tabla22 <- tabla22[with(tabla22, order(tabla22$Date)), ] 
colnames(tabla22) <- c("Date","Open","High","Low","Close","Volume","Market_Cap")
tabla22$Close<-as.numeric(gsub(",", "", tabla22$Close))
ds2 <- tabla22$Date
y2 <- log(tabla22$Close)
df2 <- data.frame(ds2, y2)
colnames(df2) <- c("ds","y")
m2 <- prophet(df2)
future2 <- make_future_dataframe(m2, periods = 365)
forecast2 <- predict(m2, future2)

plot(m2, forecast2)

```

> Fuente: Elaboración propia


### Tendencia ETH

```{r}
prophet_plot_components(m2, forecast2)
```

> Fuente: Elaboración propia

Row
-------------------------------

### Gráfico Interactivo ETH
```{r}
dyplot.prophet(m2, forecast2)
```

> Fuente: Elaboración propia

Ripple
=====================================

Row
-------------------------------------

### Pronóstico XRP

```{r}
datos3 <- read_html(fuente3) 
tabla3 <- datos3 %>%
  html_nodes("table") %>%
  html_table(header=T)
tabla32 <- tabla3[[3]]
tabla32$Date <- mdy(tabla32$Date)
tabla32 <- tabla32[with(tabla32, order(tabla32$Date)), ] 
colnames(tabla32) <- c("Date","Open","High","Low","Close","Volume","Market_Cap")
tabla32$Close<-as.numeric(gsub(",", "", tabla32$Close))
ds3 <- tabla32$Date
y3 <- log(tabla32$Close)
df3 <- data.frame(ds3, y3)
colnames(df3) <- c("ds","y")
m3 <- prophet(df3)
future3 <- make_future_dataframe(m3, periods = 365)
forecast3 <- predict(m3, future3)

plot(m3, forecast3)

```

> Fuente: Elaboración propia

### Tendencia XRP

```{r}
prophet_plot_components(m3, forecast3)
```

> Fuente: Elaboración propia

Row
-------------------------------

### Gráfico Interactivo XRP
```{r}
dyplot.prophet(m3, forecast3)
```

> Fuente: Elaboración propia

Litecoin
=====================================

Row
-------------------------------------

### Pronóstico LTC

```{r}
datos4 <- read_html(fuente4) 
tabla4 <- datos4 %>%
  html_nodes("table") %>%
  html_table(header=T)
tabla42 <- tabla4[[3]]
tabla42$Date <- mdy(tabla42$Date)
tabla42 <- tabla42[with(tabla42, order(tabla42$Date)), ] 
colnames(tabla42) <- c("Date","Open","High","Low","Close","Volume","Market_Cap")
tabla42$Close<-as.numeric(gsub(",", "", tabla42$Close))
ds4 <- tabla42$Date
y4 <- log(tabla42$Close)
df4 <- data.frame(ds4, y4)
colnames(df4) <- c("ds","y")
m4 <- prophet(df4)
future4 <- make_future_dataframe(m4, periods = 365)
forecast4 <- predict(m4, future4)

plot(m4, forecast4)

```

> Fuente: Elaboración propia

### Tendencia LTC

```{r}
prophet_plot_components(m4, forecast4)
```

> Fuente: Elaboración propia

Row
-------------------------------

### Gráfico Interactivo LTC
```{r}
dyplot.prophet(m4, forecast4)
```

> Fuente: Elaboración propia

Bitcoin Cash
=====================================

Row
-------------------------------------

### Pronóstico BCH

```{r}
datos5 <- read_html(fuente5) 
tabla5 <- datos5 %>%
  html_nodes("table") %>%
  html_table(header=T)
tabla52 <- tabla5[[3]]
tabla52$Date <- mdy(tabla52$Date)
tabla52 <- tabla52[with(tabla52, order(tabla52$Date)), ] 
colnames(tabla52) <- c("Date","Open","High","Low","Close","Volume","Market_Cap")
tabla52$Close<-as.numeric(gsub(",", "", tabla52$Close))
ds5 <- tabla52$Date
y5 <- log(tabla52$Close)
df5 <- data.frame(ds5, y5)
colnames(df5) <- c("ds","y")
m5 <- prophet(df5)
future5 <- make_future_dataframe(m5, periods = 365)
forecast5 <- predict(m5, future5)

plot(m5, forecast5)

```

> Fuente: Elaboración propia

### Tendencia BCH

```{r}
prophet_plot_components(m5, forecast5)
```

> Fuente: Elaboración propia

Row
-------------------------------

### Gráfico Interactivo BCH
```{r}
dyplot.prophet(m5, forecast5)
```

> Fuente: Elaboración propia

Basic Attention Token
=====================================

Row
-------------------------------------

### Pronóstico BAT

```{r}
datos6 <- read_html(fuente6) 
tabla6 <- datos6 %>%
  html_nodes("table") %>%
  html_table(header=T)
tabla62 <- tabla6[[3]]
tabla62$Date <- mdy(tabla62$Date)
tabla62 <- tabla62[with(tabla62, order(tabla62$Date)), ] 
colnames(tabla62) <- c("Date","Open","High","Low","Close","Volume","Market_Cap")
tabla62$Close<-as.numeric(gsub(",", "", tabla62$Close))
ds6 <- tabla62$Date
y6 <- log(tabla62$Close)
df6 <- data.frame(ds6, y6)
colnames(df6) <- c("ds","y")
m6 <- prophet(df6)
future6 <- make_future_dataframe(m6, periods = 365)
forecast6 <- predict(m6, future6)

plot(m6, forecast6)

```

> Fuente: Elaboración propia

### Tendencia BAT

```{r}
prophet_plot_components(m6, forecast6)
```

> Fuente: Elaboración propia

Row
-------------------------------

### Gráfico Interactivo BAT
```{r}
dyplot.prophet(m6, forecast6)
```

> Fuente: Elaboración propia

Decentraland
=====================================

Row
-------------------------------------

### Pronóstico MANA

```{r}
datos7 <- read_html(fuente7) 
tabla7 <- datos7 %>%
  html_nodes("table") %>%
  html_table(header=T)
tabla72 <- tabla7[[3]]
tabla72$Date <- mdy(tabla72$Date)
tabla72 <- tabla72[with(tabla72, order(tabla72$Date)), ] 
colnames(tabla72) <- c("Date","Open","High","Low","Close","Volume","Market_Cap")
tabla72$Close<-as.numeric(gsub(",", "", tabla72$Close))
ds7 <- tabla72$Date
y7 <- log(tabla72$Close)
df7 <- data.frame(ds7, y7)
colnames(df7) <- c("ds","y")
m7 <- prophet(df7)
future7 <- make_future_dataframe(m7, periods = 365)
forecast7 <- predict(m7, future7)

plot(m7, forecast7)

```

> Fuente: Elaboración propia

### Tendencia MANA

```{r}
prophet_plot_components(m7, forecast7)
```

> Fuente: Elaboración propia

Row
-------------------------------

### Gráfico Interactivo MANA
```{r}
dyplot.prophet(m7, forecast7)
```

> Fuente: Elaboración propia

Golem
=====================================

Row
-------------------------------------

### Pronóstico GNT

```{r}
datos8 <- read_html(fuente8) 
tabla8 <- datos8 %>%
  html_nodes("table") %>%
  html_table(header=T)
tabla82 <- tabla8[[3]]
tabla82$Date <- mdy(tabla82$Date)
tabla82 <- tabla82[with(tabla82, order(tabla82$Date)), ] 
colnames(tabla82) <- c("Date","Open","High","Low","Close","Volume","Market_Cap")
tabla82$Close<-as.numeric(gsub(",", "", tabla82$Close))
ds8 <- tabla82$Date
y8 <- log(tabla82$Close)
df8 <- data.frame(ds8, y8)
colnames(df8) <- c("ds","y")
m8 <- prophet(df8)
future8 <- make_future_dataframe(m8, periods = 365)
forecast8 <- predict(m8, future8)

plot(m8, forecast8)

```

> Fuente: Elaboración propia

### Tendencia GNT

```{r}
prophet_plot_components(m8, forecast8)
```

> Fuente: Elaboración propia

Row
-------------------------------

### Gráfico Interactivo GNT
```{r}
dyplot.prophet(m8, forecast8)
```

> Fuente: Elaboración propia

Resumen
===========================================

Acerca de la metodología

* La información es obtenida mediante scraping de: https://coinmarketcap.com
* Se implementa un modelo prophet simple, para series de tiempo no lineales de la forma:

  y(t) = g(t) + s(t) + h(t) + e(t)

  Donde:

  g(t) = tendencia de los modelos (modelo de crecimiento saturado y un modelo lineal por partes) que describe el aumento o la disminución a largo plazo de los datos.
  
  s(t) = modelo para la estacionalidad con la serie Fourier para describir efectos por factores estacionales como la época del año
  
  h(t) = modelo para la estacionalidad para describir efectos por factores estacionales como efectos vacacionales y eventos de gran magnitud ("human-scale” seasonalities)
  
  e(t) = el término de erreor asociado al modelo
   
* El pronóstico se realiza a 365 días a partir de `r format(Sys.Date(), format = "%B %d, %Y")`

* Para más información sobre prophet

  https://towardsdatascience.com/forecasting-with-prophet-d50bbfe95f91
  
  https://research.fb.com/blog/2017/02/prophet-forecasting-at-scale/
  
  https://github.com/facebook/prophet


Créditos
========================================

Created by: Cesar Hernández

Correo: cghv94@outlook.es

Twitter: @[cghv94](https://twitter.com/cghv94)

>
Esta obra fue generada mediante R en `r format(Sys.Date(), format = "%B %d, %Y")`. [Creative Commons (CC)](http://creativecommons.org/licenses/by-nc-sa/4.0/).
Esta obra está bajo una licencia de Creative Commons Reconocimiento-NoComercial-CompartirIgual 4.0 Internacional.