watervogels.Rmd

---
title: "Watervogels"
author: "gehau"
date: "December 7, 2017"
output:
  html_document: default
  pdf_document: default
---

```{r setup, include=FALSE}

knitr::opts_chunk$set(echo = TRUE)
library(tidyverse)


```

## Watervogels

```{r lees_bestand}
setwd("/home/gehau/git/watervogels")
waarnemingen <- readRDS("data/V3/occurrence.rds")
waarnemingen <-  waarnemingen[which( !is.na(waarnemingen$decimalLatitude), arr.ind=TRUE),]
waarnemingen <-  waarnemingen[which( !is.na(waarnemingen$decimalLongitude), arr.ind=TRUE),]


```

```{r reshape , include=FALSE}
require(reshape) #om colsplit te gebruiken
```

```{r transform , warning=FALSE}
date_split <-
  colsplit(waarnemingen$eventDate,
           split = "-",
           names = c('jaar', 'maand', 'tmp'))
day_split <-
  colsplit(date_split$tmp,
           split = "T",
           names = c('dag', 'tijd'))
jaar <- as.character(date_split$jaar)
maand <- as.character(date_split$maand)
dag <- as.character(day_split$dag)
tijd <- as.character(day_split$tijd)
soort <- waarnemingen$vernacularName
aantal_individuen <- waarnemingen$individualCount
waarnemingen <- cbind(waarnemingen, soort, aantal_individuen, jaar, maand, dag, tijd)

wnm <- subset(waarnemingen, select = c(eventID,soort, aantal_individuen,jaar, maand, dag, tijd))

```

```{r ggplot, warning=FALSE}
# create plot
    p <- ggplot(data = wnm, aes(x = aantal_individuen))  + theme(axis.text.x = element_text(angle = 60, hjust = 1))
    
    # add geom
    p <- p + geom_histogram() 
    
    p <- p + scale_x_log10() 
    
    p <- p + aes(fill = maand) 
    
    p <- p + scale_fill_discrete(
      breaks=c("1","2","3","4","5","6","7","8","9","10","11","12"), 
      labels=c("januari", "februari", "maart", "april", "mei", "juni", "juli", "augustus", "september", "oktober", "november", "december")) 
    
    p
```


```{r map, warning=FALSE}

    p <- ggplot(data = wnm, aes(x = jaar, y = aantal_individuen)) + theme(axis.text.x = element_text(angle = 60, hjust = 1))
    
    # add geom
    p <- p + geom_bar(stat = 'identity', width = .5)
    
    # update data layer with new mapping
    p <- p %+% aes(fill = maand)
    
     p <- p + scale_fill_discrete(
      breaks=c("1","2","3","4","5","6","7","8","9","10","11","12"), 
      labels=c("januari", "februari", "maart", "april", "mei", "juni", "juli", "augustus", "september", "oktober", "november", "december")) 
    
    
    # titel en subtitel
    p <-  p %+% labs(subtitle = "todo",  title = "Aantal watervogels")
    p
```


```{r ggplot2, warning=FALSE}

# add facets + fix x-axis
p + facet_wrap(~maand, scales = 'free_x')

```


```{r boxplot}

# create boxplot var
bp <- ggplot(wnm, aes(x = jaar, y = aantal_individuen, fill = maand)) +
  geom_boxplot() + scale_y_log10() + theme(axis.text.x = element_text(angle = 60, hjust = 1)) + 
  scale_fill_discrete(
      breaks=c("1","2","3","4","5","6","7","8","9","10","11","12"), 
      labels=c("januari", "februari", "maart", "april", "mei", "juni", "juli", "augustus", "september", "oktober", "november", "december")) + labs(subtitle = "todo",  title = "Aantal watervogels") 
# boxplot
bp

# we'll use this a lot  
fw <- facet_wrap(~maand, scales = 'free_y')

bp + fw
# 
# bp.h <- bp %+% aes(x = as.factor(jaar)) 
# 
# bp.h + fw


```

# Create correlation grid

```{r dplyr , include=FALSE}
library(dplyr) #
```
Top tien van de watervogels
```{r topTien, warning=FALSE}
vogelsoorten <- subset(wnm, select = c(soort, aantal_individuen))
vogelsoorten  %>% 
  group_by(soort) %>% 
  summarise(aantal_individuen = sum(aantal_individuen)) %>%
  arrange(desc(aantal_individuen)) %>% 
  head(n = 10L)
```

How correlated is the abundance of different species?
```{r GGally , include=FALSE}
# import library
library(GGally) 
```

```{r corr, warning=FALSE}
# create a sample of 1000 obs

sample_wnm <- subset(wnm, select = c(eventID, soort, aantal_individuen, maand)) %>% 
  spread(soort, aantal_individuen)%>% #spread dplyr
  subset(select = c(Smient,Kievit,Kokmeeuw,Kuifeend,Tafeleend,Kolgans,maand))
  
set.seed(888) 
#sample_wnm <- sample_wnm[sample(1:nrow(sample_wnm),1000),]


# look at correlations
ggpairs(data=sample_wnm, # data.frame with variables
        title="Bird species occurence correlations") + theme(axis.text.x = element_text(angle = 60, hjust = 1)) 
```

```{r add alpha and trendline, warning=FALSE}

ggpairs(data = sample_wnm, 
        lower = list(continuous = wrap("smooth", alpha=1/5, shape = I('.'), 
                                       colour ='blue')),
         title="Bird species occurence correlations") + theme(axis.text.x = element_text(angle = 60, hjust = 1))




```


```{r line plot}


# create line plot
p <- ggplot( waarnemingen[waarnemingen$municipality == "Kalmthout",], aes(verbatimEventDate, individualCount)) +
              geom_line(colour = 'darkgrey') + scale_y_log10() #+ theme(axis.text.x = element_text(angle = 60, hjust = 1))


# change default grouping
p <- p %+% aes(group = vernacularName)
p + facet_wrap(~jaar, scales = 'free_x')


```