03_exercise1.qmd

::: {.content-hidden unless-format="html"}
$$
\newcommand{\Exp}{\mathrm{E}}
\newcommand\given[1][]{\:#1\vert\:}
\newcommand{\Cov}{\mathrm{Cov}}
\newcommand{\Var}{\mathrm{Var}}
\newcommand{\rank}{\mathrm{rank}}
\newcommand{\bm}[1]{\boldsymbol{\mathbf{#1}}}
$$
:::


# Exercises I

### Required packages {.unnumbered}

```{r, message = FALSE, warning = FALSE, results = 'hide'}
pkgs <- c("sf", "mapview", "areal", "spdep", "spatialreg", "tmap", "viridisLite",
          "ggplot2", "ggthemes", "gridExtra") # note: load spdep first, then spatialreg
lapply(pkgs, require, character.only = TRUE)

```

### Session info {.unnumbered}

```{r}
sessionInfo()

```

### Reload data from pervious session {.unnumbered}

```{r}
load("_data/msoa2_spatial.RData")
```



## General Exercises

### 1) Can you import the spatial administrative units of Germany ("Kreisgrenzen_2020_mit_Einwohnerzahl" in _data folder) and make a simple plot of the boundaries?   {.unnumbered}

```{r}
# Import shape file layer
ger.sdpf <- st_read(dsn = "_data/Kreisgrenzen_2020_mit_Einwohnerzahl",
                    layer = "KRS_ew_20")

# Plot via ggplot
gp <- ggplot(ger.sdpf)+
    geom_sf( color = "magenta", fill = NA)+
    coord_sf(datum = NA)+
    theme_map()
gp
```

### 2) What is the Coordinate reference system of this German shape file? {.unnumbered}

```{r}
st_crs(ger.sdpf)
```


### 3) Using the MSOA data in London from above: Please draw a map showing the distribution of share of non-EU immigrants (per_nonEU).  {.unnumbered}

```{r}
gp1 <- ggplot(msoa.spdf)+
    geom_sf(aes(fill = per_nonEU))+
    scale_fill_viridis_c(option = "A")+
    coord_sf(datum = NA)+
    theme_map()+
    theme(legend.position = c(.9, .4))
gp1
```

### 4) Below we will load some additional data on heat islands.   {.unnumbered}

We will add some Data about [London's Urban Heat Island](https://data.london.gov.uk/dataset/london-s-urban-heat-island---average-summer/). It contains information about the mean temperature at midnight during the summer of 2011.

This is a tif file that we need to read in with stars and then transform into sf.

```{r}
library(stars)

# Read geo_tif with stars
urbclim <- read_stars("https://data.london.gov.uk/download/ae16d5af-5dce-49bc-b1e2-88bb41e8bfd0/f4e3a05d-fad7-4b56-8c42-ba2274f3bb3a/London_Tmin_midnight_2011.tif")

# urbclim <- read_stars("_data/London_Tmin_midnight_2011.tif")

# Transfer to sf
urbclim.spdf <- st_as_sf(urbclim)
names(urbclim.spdf)[1] <- "Tmin_midnight"
```

a) On which projection is the urbclim temperature data?

b) Can you please calculate the average night-time temperature for each MSOA using area weighted interpolation. Make sure that the objects are on the same projections / crs.

c) Create a map showing the temperature for each MSOA, and plot it next to the maps of non-EU immigrant residents (e.g. using `grid.arrange()`) .


```{r}
# Check projection
st_crs(urbclim.spdf)
```


```{r}
# Add id
urbclim.spdf$id <- rownames(urbclim.spdf)

# Bring on common crs
urbclim.spdf <- st_transform(urbclim.spdf, crs = st_crs(msoa.spdf))

# Use area weights interpolation to merge
msoa.spdf <- aw_interpolate(
  msoa.spdf,
  tid = "MSOA11CD",
  source = urbclim.spdf,
  sid = "id",
  weight = "sum",
  output = "sf",
  intensive = "Tmin_midnight"             
)
```




```{r}
# Make map of Temperature
gp2 <- ggplot(msoa.spdf)+
    geom_sf(aes(fill = Tmin_midnight))+
    scale_fill_viridis_c(option = "C")+
    coord_sf(datum = NA)+
    theme_map()+
    theme(legend.position = c(.9, .4))

# Plot two ggplot maps next to each other
gp <- grid.arrange(gp1, gp2, nrow = 1)
gp
```


### 4) What's the correlation between the share of non-Eu residents and the termperature.   {.unnumbered}

```{r}
mod1.lm <- lm(Tmin_midnight ~ per_nonEU, data = msoa.spdf)
summary(mod1.lm)
```