Airbnb European Snapshot

ggplot2

Parameterized HTML dashboards for various cities

Published

April 11, 2023

Starting with the datasets

The original datasets were all based on city and listing type (weekday or weekend). I needed to clean the data a bit and wrangle them into one main dataset. The parameter of interest was city.

Behind the scenes

library(tidyverse)
library(ggplot2)
library(dplyr)
library(purrr)
library(data.table)
library(gsubfn)
library(tidyr)
library(flexdashboard)


files <- dir(pattern = "*.csv")
file.list <- sapply(files, read.csv, simplify=FALSE)
airbnb <- rbindlist(file.list, idcol="city")[, city := gsub("(.*)[_]([^.]+)[.].*", "\\1 \\2",city)]
airbnb<-separate(airbnb, city, into = c("city", "type"), sep = " (?=[^ ]+$)")
write.csv(airbnb, "airbnb.csv", row.names=FALSE)

Dashboard pieces

Next up was to subset the data based on city and create a color palette for the dashboard based on Airbnb brand colors. Then the various plots used in the dashboard were created.

Behind the scenes

airbnb_subset <- airbnb %>%
  filter(city == params$city)

cbPalette <- c("#5AB2FF", "#00A699", "#FFE25A", "#845AFF", "#FC642D", "#FF5A5F")

bedplot<-airbnb_subset %>%
  ggplot(aes(bedrooms, fill=type)) +
    geom_bar(stat="count",position="dodge")+
    theme_minimal()+
    theme(legend.position="bottom")+ 
    labs(fill="", x = "Number of Bedrooms", y = "")  +
    scale_fill_manual(values=cbPalette)


cleanplot<-airbnb_subset %>%
  ggplot(aes(cleanliness_rating, fill=room_type)) +
    geom_bar(stat="count",position="dodge")+
    theme_minimal()+
    theme(legend.position="bottom")+ 
    labs(fill="", x = "Cleanliness Rating", y = "")  +
    scale_fill_manual(values=cbPalette)

dist<-airbnb_subset %>%
  filter(realSum<4000) %>%
  ggplot(aes(x=metro_dist, y=realSum, color=type)) +
    geom_point(alpha=.5, size=2)+
    theme_minimal()+
    theme(legend.position="bottom")+ 
    labs(color="", x = "Distance from City Center", y = "Total Price") +
    scale_color_manual(values=cbPalette)

Dashboard layout

Finally the final layout was created to show the valueboxes and the plots that were layed out above.

Behind the scenes

articles <- nrow(airbnb_subset)
valueBox(articles, icon = "fa-home")

super <- nrow(airbnb_subset %>% filter(host_is_superhost == "True"))
valueBox(super, icon = "fa-user")

avgprice <- round(mean(airbnb_subset$realSum),digits = 0)
valueBox(avgprice, icon = "fa-coins")

avgprice <- round(mean(airbnb_subset$guest_satisfaction_overall),digits = 0)
valueBox(avgprice, icon = "fa-star")

bedplot

cleanplot

dist

Rendering code

In a separate render.r file the following code was run, resulting in the various HTML dashboards.

Behind the scenes

city <- unique(airbnb$city)

reports <- tibble(
  output_file = stringr::str_c("output/", city, "-airbnb.html"),
  params = map(city, ~list(city = .))
)

reports %>%
  pwalk(rmarkdown::render, input = "airbnb/parameter.qmd")

Final HTML Dashboards

Here are the final HTML dashboards that are produced:

« Back to Projects