index.Rmd

---
title: "R demo WEDC Master's students"
author: "A.Khouakhi"
date: "December 1, 2017"
output:
  html_document:
    highlight: haddock
    theme: readable
  pdf_document: default
  word_document: default
---

### 1. Load the necessary libraries

```{r, eval=T, echo=T,include=T,warning= F, message=F,cache=FALSE}

library(tidyverse)
library(GGally)
library(tibbletime)
library(kableExtra)
library(lubridate)
library(leaflet)
```

### 2. Load and read the data

The data contains some rows that we don't need so we will skip them using the argument `skip`

```{r, eval=T, echo=T,warning= F, message=F,cache=FALSE}
# Data path
data_path <- "water_level.csv"
# Read the data 
stage <- read_csv(data_path, skip = 21, progress = T) #skip = 21
stage 
```
### 3. Tidying up the data 

The variable names are not consistent so we are going to rename them using the rename function from the `dplyr` package  

```{r, eval=T, echo=T}
# Change variable names 
stage <- rename(stage, st.4082 = s_15_4082.csv,
         st.4086 = s_15_4086.csv,
         st.4145 = s_4145.csv,
         st.4161 = s_4161.csv, 
         st.4174 = s_4174.csv,
         st.4195 = s_4195.csv)
```

### 4. Analyses 
As example we are going to compute the annual maximum water level for each gauging station and check if there is any trend

```{r, eval=T, echo=T,warning= F, message=F,cache=FALSE}
# Compute the annual maximum water level at each gauging station
stage <- as_tbl_time(stage, date_time)

ann_max <- stage %>% 
  mutate(year =year(date_time)) %>% 
  group_by(year) %>% 
  summarise_all(funs(max))




knitr::kable(ann_max, "html") %>%
  kable_styling(bootstrap_options = c("striped", "hover", "condensed"))
```

### 5. Modelling and visualization 
fit a linear regression model to the data, see if there is any correlation between different stations and so on

```{r, eval=T, echo=T}

# Plot the data and and fit linear regression line
plot1 <- ann_max %>% gather("station", "stage", 3:8) %>% 
  ggplot(., aes(date_time, stage))+
  theme_bw()+
  geom_point()+
  geom_smooth(method = "lm", se=T, color="blue")+
  facet_grid(station ~ ., scales = "free_y") 
plot1
# Add lines
plot1 + geom_line()


# Plot in small graphics
plot1 + facet_wrap(~station)

# Summary graphic 
ggpairs(ann_max, columns = 3:8)

# Correlation matrix 
ggcorr(ann_max[, -1],label = TRUE,
       label_alpha = TRUE)

```

### 6. Other visualization 


```{r, eval=T, echo=T}

# Plot in the same graphic   
  clrs <-  colorRamps::primary.colors(6)
  ann_max %>% gather("station", "stage", 3:8) %>% 
    ggplot(.,aes(x = date_time, y = stage, colour = station))+
    theme_bw()+
    geom_line()+
    geom_point()+
    scale_color_manual(values = clrs)
    

```

### 7. Mapping gauge locations
Mapping station locations using interactive leaflet map 

```{r, eval=T, echo=T,warning= F, message=F,cache=FALSE}

# Read gauges coordinates 
locations <- read_csv("gauges_coord.csv")

leaflet(data = locations) %>% addTiles() %>%
  addMarkers(~long, ~lat, popup = ~as.character(riverName),
             label = ~as.character(riverName))
    
```