Skip to content

Latest commit

 

History

History
164 lines (133 loc) · 6.19 KB

assignment4solution.md

File metadata and controls

164 lines (133 loc) · 6.19 KB

Statistical assignment 4

[add your name here] [add your candidate number here - mandatory] [add date here]

In this assignment you will need to reproduce 5 ggplot graphs. I supply graphs as images; you need to write the ggplot2 code to reproduce them and knit and submit a Markdown document with the reproduced graphs (as well as your .Rmd file).

First we will need to open and recode the data. I supply the code for this; you only need to change the file paths.

```r
library(tidyverse)
Data8 <- read_tsv("C:\\Users\\ab789\\datan3_2019\\data\\UKDA-6614-tab\\tab\\ukhls_w8\\h_indresp.tab")

Data8 <- Data8 %>%
    select(pidp, h_age_dv, h_payn_dv, h_gor_dv)

Stable <- read_tsv("C:\\Users\\ab789\\datan3_2019\\data\\UKDA-6614-tab\\tab\\ukhls_wx\\xwavedat.tab")

Stable <- Stable %>%
    select(pidp, sex_dv, ukborn, plbornc)

Data <- Data8 %>% left_join(Stable, "pidp")

rm(Data8, Stable)

Data <- Data %>%
    mutate(sex_dv = ifelse(sex_dv == 1, "male",
                       ifelse(sex_dv == 2, "female", NA))) %>%
    mutate(h_payn_dv = ifelse(h_payn_dv < 0, NA, h_payn_dv)) %>%
    mutate(h_gor_dv = recode(h_gor_dv,
                     `-9` = NA_character_,
                     `1` = "North East",
                     `2` = "North West",
                     `3` = "Yorkshire",
                     `4` = "East Midlands",
                     `5` = "West Midlands",
                     `6` = "East of England",
                     `7` = "London",
                     `8` = "South East",
                     `9` = "South West",
                     `10` = "Wales",
                     `11` = "Scotland",
                     `12` = "Northern Ireland")) %>%
    mutate(placeBorn = case_when(
            ukborn  == -9 ~ NA_character_,
            ukborn < 5 ~ "UK",
            plbornc == 5 ~ "Ireland",
            plbornc == 18 ~ "India",
            plbornc == 19 ~ "Pakistan",
            plbornc == 20 ~ "Bangladesh",
            plbornc == 10 ~ "Poland",
            plbornc == 27 ~ "Jamaica",
            plbornc == 24 ~ "Nigeria",
            TRUE ~ "other")
    )
```

Reproduce the following graphs as close as you can. For each graph, write two sentences (not more!) describing its main message.

  1. Histogram (20 points)

    Data %>%
    ggplot(aes(x = h_age_dv)) +
    geom_histogram(binwidth = 1) +
    xlab("Age") +
    ylab("Number of respondents")

  2. Scatter plot (20 points). The red line shows a linear fit; the blue line shows a quadratic fit. Note the size and position of points.

    Data %>% 
        ggplot(aes(x = h_age_dv, y = h_payn_dv)) +
        geom_point(size = 0.1, position = "jitter") +
        geom_smooth(method = "lm", colour = "red") +
        geom_smooth(method = "lm", formula = y ~ x + I(x^2), colour = "blue") +
        xlim(16, 65) +
        xlab("Age") +
        ylab("Monthly earnings")

  3. Faceted density chart (20 points).

    Data %>%
    filter(!is.na(placeBorn)) %>%
    ggplot(aes(x = h_age_dv)) +
    geom_density(fill = "black") +
    xlab("Age") +
    facet_wrap(~ placeBorn)

  4. Ordered bar chart of summary statistics (20 points).

    Data %>%
    filter(!is.na(placeBorn)) %>%
    filter(!is.na(sex_dv)) %>%
    group_by(placeBorn, sex_dv) %>%
    summarise(
            medianIncome = median(h_payn_dv, na.rm = TRUE)
    ) %>%
    ggplot(aes(x = reorder(placeBorn, medianIncome), y = medianIncome, fill = sex_dv)) +
    geom_bar(stat = "identity", position = "dodge") +
    scale_fill_manual(values = c("darkred", "darkblue")) +
    coord_flip() +
    ylab("Median net monthly earnings") +
    xlab("Country of birth") +
    theme(legend.position="top") +
    guides(fill=guide_legend(reverse=TRUE)) +
    labs(fill="")

  5. Map (20 points). This is the most difficult problem in this set. You will need to use the NUTS Level 1 shape file (available here -- https://data.gov.uk/dataset/2aa6727d-c5f0-462a-a367-904c750bbb34/nuts-level-1-january-2018-full-clipped-boundaries-in-the-united-kingdom) and a number of packages for producing maps from shape files. You will need to google additional information; there are multiple webpages with the code that produces similar maps.

    library(rgdal)
library(ggmap)
library(mapproj)
library(rgeos)

shapefile <- readOGR(dsn = getwd(), layer = "NUTS_Level_1_January_2018_Full_Clipped_Boundaries_in_the_United_Kingdom", verbose = FALSE)
mapdata <- broom::tidy(shapefile, region="nuts118nm")

mapdata <- mapdata %>%
    mutate(id = recode(id,
                       "East Midlands (England)" = "East Midlands",
                       "North East (England)" = "North East",
                       "North West (England)" = "North West",
                       "South East (England)" = "South East",
                       "South West (England)" = "South West",
                       "West Midlands (England)" = "West Midlands",
                       "Yorkshire and The Humber" = "Yorkshire"))

medianIncome <- Data %>%
    filter(!is.na(sex_dv)) %>%
    group_by(h_gor_dv) %>%
    summarise(
            medianIncome = median(h_payn_dv, na.rm = TRUE)
    )

mapdata2 <- mapdata %>%
    left_join(medianIncome, by = c("id" = "h_gor_dv"))


ggplot() +
    geom_polygon(data = mapdata2, aes(x = long, y = lat, group = group, fill = medianIncome)) +
    scale_fill_gradient(trans = "reverse") +
    labs(fill="") +
    theme_void() +
    coord_map() +
    ggtitle("Median earnings by region (£)")