---
title: 'Session 03 : Data Visualization'
author: '<your-name>'
output: html_document
---

== Task 1 : Configure R and load data

=== Install relevant packages


```{r, configure-r, eval=FALSE, echo=TRUE}
# Load required packages. Note that the install() commands are commented out as packages are already downloaded on lab machines
# install.packages("tidyverse")
library(tidyverse)
# install.packages("sf") # SimpleFeatures package for working with spatial data.
library(sf)
# install.packages("tmap") # tmap library, which uses syntax very similar to ggplot2.
library(tmap)
```

=== Read in the dataset

```{r, download-data, eval=FALSE, echo=TRUE}
# A pre-prepared simple features data frame that loads Local Authority outlines
# (as sfc_MULTIPOLYGON) and 2011 Census data that we'll use
# to model the vote for the remaining sessions.
data_gb <- st_read("http://homepages.see.leeds.ac.uk/~georjb/geocomp/03_datavis/data/data_gb.geojson")
```

== Task 2. Perform some exploratory data analysis

=== Do some data wrangling


```{r, data-wrangle, eval=FALSE, echo=TRUE}
# Calculate the LA share of Leave vote by Region.
region_summary <- data_gb %>%
      group_by(Region) %>%
        summarise(share_leave=sum(Leave)/sum(Leave, Remain)) %>%
           arrange(desc(share_leave))
# Clear geometry data.
st_geometry(region_summary) <- NULL
print(region_summary)
```

=== Create some summary graphics

```{r, summary-graphics, eval=FALSE, echo=TRUE}
# Create bar chart of result ordered by LA.
data_gb %>%
  ggplot(aes(x=reorder(lad15nm,-share_leave), y=margin_leave, fill=margin_leave))+
  geom_bar(stat="identity", width=1)+
  scale_fill_distiller(palette = 5, type="div", direction=1, guide="colourbar", limits=c(-0.3,0.3))+
  scale_x_discrete(breaks=c("Lambeth","Slough","Boston")) +
  geom_hline(aes(yintercept=0))+
  theme_classic()+
  xlab("LAs by Leave (asc)")+
  ylab("Margin Leave/Remain")
```


== Task 3. Perform some exploratory spatial data analysis

=== Explore spatial variation in the Leave:Remain vote

```{r, choropleth-leave, eval=FALSE, echo=TRUE}
# Generate a choropleth displaying share of leave vote by LAD.
tm_shape(data_gb) +
  tm_fill(col="share_leave", style="cont", size=0.2, id="geo_label", palette="Blues", title="") +
  tm_borders(col="#bdbdbd", lwd=0.5) +
  tm_layout(
    title="LA share of Leave vote",
    title.snap.to.legend=TRUE,
    title.size=0.8,
    legend.text.size=0.6,
    title.position = c("right", "center"),
    legend.position = c("right","center"),
    frame=FALSE,
    legend.outside=TRUE)
```

== Task 4. Explore bivariate associations using correlation

```{r, scatterplot, eval=FALSE, echo=TRUE}
# Calculate correlation coefficient of share Leave by degree-educated.
data_gb %>%
  summarise(cor(share_leave, degree_educated))

# Generate scatterplot of share Leave by degree-educated.
data_gb %>%
  ggplot(aes(x=share_leave, y=degree_educated)) +
  geom_point(colour="#525252",pch=21, alpha=0.8) +
  theme_bw()
```

== Task 5. Generate graphical small multiples

```{r, small-multiples, eval=FALSE, echo=TRUE}
data_gb %>%
  gather(c(younger_adults:eu_born), key = "expl_var", value="la_prop") %>%
  ggplot(aes(x=la_prop, y=share_leave))+
  geom_point(colour="#525252",pch=21)+
  facet_wrap(~expl_var, scales="free")+
  theme_bw()
```