---
title: "Pronostico Aceite de Soja (ZL=F)"
output: 
  flexdashboard::flex_dashboard:
    orientation: columns
    vertical_layout: fill
    source_code: embed
    theme: flatly
---
  
```{r setup, include=FALSE}

#install.packages("quantmod")
#install.packages("ggplot2")
#install.packages("forecast")
#install.packages("nnfor")
#install.packages("plotly")

library(flexdashboard)
library(quantmod) #para importar la serie
library(ggplot2)
library(forecast)
library(nnfor) #redes neuronales
library(plotly)


soja.df=getSymbols("ZL=F",
                   from="2016-01-01",
                   to="2023-04-01",
                   env=NULL,
                   periodicity ="weekly",
                   return.class="data.frame")

soja.df$Fecha=as.Date(row.names(soja.df),format="%Y-%m-%d")
#row.names(soja.df)=NULL
names(soja.df)= c("Apertura","Maximo","Minimo","Cierre","Volumen","Ajustado","Fecha")

sum(is.na(soja.df))
soja.df=na.omit(soja.df)

soja.ts=ts(soja.df$Ajustado,start = c(2016,1),frequency = 52)

train=window(soja.ts,end=c(2022,52))
test=window(soja.ts,start=c(2023,1))
n=length(test)

m=dim(soja.df)[1]
f1=soja.df[(m-n+1):m,7]
f2=soja.df[m,7]+7*seq(1:15)
```

Serie
=======================================================================
Column {data-width=50}
-----------------------------------------------------------------------
### Apertura
```{r}
valueBox(value = round(soja.df[dim(soja.df)[1],]$Apertura,2), 
         caption = "Apertura", 
         icon = "fa-line-chart", 
         color = "#3cb371")
```


### Maximo
```{r}
valueBox(value = round(soja.df[dim(soja.df)[1],]$Maximo,2), 
         caption = "Maximo", 
         icon = "fa-line-chart", 
         color = "#3cb371")
```


### Minimo
```{r}
valueBox(value = round(soja.df[dim(soja.df)[1],]$Minimo,2), 
         caption = "Minimo", 
         icon = "fa-line-chart", 
         color = "#3cb371")
```


### Cierre
```{r}
valueBox(value = round(soja.df[dim(soja.df)[1],]$Cierre,2), 
         caption = "Cierre", 
         icon = "fa-line-chart", 
         color = "#3cb371")
```


### Volumen
```{r}
valueBox(value = paste(soja.df[dim(soja.df)[1],]$Volumen,"K"), 
         caption = "Volumen", 
         icon = "fa-line-chart", 
         color = "#3cb371")
```


### Ajustad0
```{r}
valueBox(value = round(soja.df[dim(soja.df)[1],]$Ajustado,2), 
         caption = "Volumen", 
         icon = "fa-line-chart", 
         color = "#3cb371")
```


### Fecha
```{r}
valueBox(value = soja.df[dim(soja.df)[1],]$Fecha,
         caption = "Ultima actualizacion", 
         icon = "fa-line-chart", 
         color = "#3cb371")
```



Column {data-width=950, .tabset}
-----------------------------------------------------------------------
### Precio Ajustado
```{r,fig.width=10, fig.height=6}
#ggplot(soja.df,aes(Fecha,Ajustado))+
#  geom_line(color = "#FC4E07")+
#  ggtitle("Precios de Futuros de Aceite de Soja")+
#  xlab("")+
#  ylab("")+
#  scale_x_date(date_breaks = "1 year", date_labels = "%Y")

fig <- plot_ly(soja.df, type = 'scatter', mode = 'lines')%>%
  add_trace(x = ~Fecha, y = ~Ajustado)%>%
  layout(showlegend = F,
         xaxis = list(rangeslider = list(visible = T),
                      rangeselector=list(
                        buttons=list(
                          list(count=1,
                               label="1m",
                               step="month",
                               stepmode="backward"),
                          list(count=6,
                               label="6m",
                               step="month",
                               stepmode="backward"),
                          list(count=1,
                               label="YTD",
                               step="year",
                               stepmode="todate"),
                          list(count=1,
                               label="1y",
                               step="year",
                               stepmode="backward"),
                          list(step="all")))))

#fig <- fig %>%
#  layout(
#    xaxis = list(zerolinecolor = '#ffff',
#                 zerolinewidth = 2,
#                 gridcolor = 'ffff'),
#    yaxis = list(zerolinecolor = '#ffff',
#                 zerolinewidth = 2,
#                 gridcolor = 'ffff'),
#    plot_bgcolor='#e5ecf6', width = 900)

fig

```


### Gráfico de velas
```{r,fig.width=10, fig.height=6}
fig <- soja.df %>% plot_ly(x = ~Fecha,
                           type="candlestick",
                           open = ~Apertura,
                           close = ~Cierre,
                           high = ~Maximo,
                           low = ~Minimo) 

fig
```


### Gráfico OHLC
```{r,fig.width=10, fig.height=6}
fig <- soja.df %>% plot_ly(x = ~Fecha,
                           type="ohlc",
                           open = ~Apertura,
                           close = ~Cierre,
                           high = ~Maximo,
                           low = ~Minimo) 

fig
```

### Base de datos
```{r,fig.width=10, fig.height=6}
knitr::kable(soja.df[,1:6])
```


Descomposicion
=======================================================================
Column{.tabset}
-----------------------------------------------------------------------
### Descomposicion
```{r,fig.width=7, fig.height=6}
ggplotly(forecast::autoplot(decompose(soja.ts,type = "multiplicative")))
```

### Grafico de estacionalidad
```{r,fig.width=7, fig.height=6}
ggplotly(ggseasonplot(soja.ts))
```


Modelos
=======================================================================
Column {data-width=700, .tabset}
-----------------------------------------------------------------------
### Arima
```{r,fig.width=10, fig.height=6}
am=auto.arima(train,
              stepwise=FALSE,
              allowdrift = FALSE)
m1=forecast::forecast(am,h=n)

forecast::autoplot(m1)+
  forecast::autolayer(fitted(m1),series="ARIMA")+
  xlim(2020,2023.5)+
  ylab("")
```


### Redes Neuronales
```{r,fig.width=10, fig.height=6}
rn=nnfor::mlp(train)
m2=forecast::forecast(rn,h=n)

forecast::autoplot(m2)+
  forecast::autolayer(rn$fitted,series="RN")+
  xlim(2020,2023.5)+
  ylab("")
```


### Holt-Winters
```{r,fig.width=10, fig.height=6}
hw=HoltWinters(train,seasonal = "multiplicative")
m3=forecast::forecast(hw,h=n)

forecast::autoplot(m3)+
  forecast::autolayer(fitted(m3),series="HW")+
  xlim(2020,2023.5)+
  ylab("")
```


### Comparacion
```{r,fig.width=10, fig.height=6}
forecast::autoplot(soja.ts)+
  forecast::autolayer(m1,series="Arima",PI=FALSE)+
  forecast::autolayer(m2,series="RN",PI=FALSE)+
  forecast::autolayer(m3,series="HW",PI=FALSE)+
  xlim(2020,2023.5)+
  ylab("")
```


Column {data-width=300, .tabset}
-----------------------------------------------------------------------
### Ajustes
```{r}
comp=round(data.frame(data.frame(m1)[1],data.frame(m2)[1],data.frame(m3)[1],data.frame(test)),2)
rownames(comp)=f1
names(comp)=c("ARIMA","RN","HW","REAL")

knitr::kable(comp)
```


### Errores
```{r}
errores=round(data.frame(
  "ARIMA"=t(accuracy(m1,test))[1:5,2],
  "RN"=t(accuracy(m2,test))[1:5,2],
  "HW"=t(accuracy(m3,test))[1:5,2]),2)

knitr::kable(errores)
```



Pronostico
=======================================================================
Column {data-width=700}
-----------------------------------------------------------------------
### Gráfico
```{r,fig.width=10, fig.height=6}
mod=HoltWinters(soja.ts,seasonal = "multiplicative")
pron=forecast::forecast(mod,h=n)

forecast::autoplot(pron)+
  forecast::autolayer(fitted(pron),series="HW")+
  xlim(2020,2023.5)+
  ylab("")
```


Column {data-width=300}
-----------------------------------------------------------------------
### Tabla
```{r}
pron.df=round(data.frame(pron),2)
rownames(pron.df)=f2
names(pron.df)=c("Pron", "Lo.80", "Hi.80", "Lo.95", "Hi.95")

knitr::kable(pron.df)
```