AirBnB Los Angeles Visualizations in R
Date: Oct. 28, 2018
Visuals should be created in R with ggplot2; Provide all relevant code, comments and interpretive insights in an R Notebook; Two or more of your five visuals must somehow highlight, contest, incorporate, address or be related to, two of Rosling’s instincts that we have discussed (i.e., Gap, Negativity, Straight Line, Fear or Size). It is up to you how you want to address/incorporate/dispel these instincts in your visuals.
1- First, upload the proper package!
library(tidyverse)## ── Attaching packages ────────────────────────────────────────────────── tidyverse 1.2.1 ──## ✔ ggplot2 2.2.1 ✔ purrr 0.2.4
## ✔ tibble 1.4.2 ✔ dplyr 0.7.4
## ✔ tidyr 0.8.0 ✔ stringr 1.2.0
## ✔ readr 1.1.1 ✔ forcats 0.2.0## ── Conflicts ───────────────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()library(RColorBrewer)
library(ggmap)
library(caret)## Warning: package 'caret' was built under R version 3.4.4## Loading required package: lattice##
## Attaching package: 'caret'## The following object is masked from 'package:purrr':
##
## lift2 - Then you upload your data, which in this case is Airbnb data for Los Angeles.
AirbnbLA <- read_csv("~/Documents/Classes(School)/2018-19/Geospatial Visualization/LA Airbnb Exercise/listings.csv", na = "..")## Parsed with column specification:
## cols(
## .default = col_character(),
## id = col_integer(),
## scrape_id = col_double(),
## last_scraped = col_date(format = ""),
## host_id = col_integer(),
## host_since = col_date(format = ""),
## host_listings_count = col_integer(),
## host_total_listings_count = col_integer(),
## latitude = col_double(),
## longitude = col_double(),
## accommodates = col_integer(),
## bathrooms = col_double(),
## bedrooms = col_integer(),
## beds = col_integer(),
## square_feet = col_integer(),
## guests_included = col_integer(),
## minimum_nights = col_integer(),
## maximum_nights = col_integer(),
## availability_30 = col_integer(),
## availability_60 = col_integer(),
## availability_90 = col_integer()
## # ... with 14 more columns
## )## See spec(...) for full column specifications.Well this is messy. I definitely can’t use some of these variables, like name, url, reviews, etc.. Let’s clean it up.
head(AirbnbLA)3 - Next, clean up the code.
Los_Angeles <- AirbnbLA %>% filter(city == "Los Angeles") %>% na.omit() %>% select(city, cleaning_fee, price, property_type, room_type, accommodates, host_since, host_is_superhost, weekly_price, monthly_price, minimum_nights, number_of_reviews, review_scores_rating, review_scores_accuracy, review_scores_cleanliness, review_scores_communication, review_scores_location, latitude, longitude)
Los_Angeles$price_2 = as.numeric(gsub("\\$", "", Los_Angeles$price))## Warning: NAs introduced by coercionLos_Angeles$cleaning_fee_2 = as.numeric(gsub("\\$", "", Los_Angeles$cleaning_fee))
Los_Angeles$weekly_price_2 = as.numeric(gsub("\\$", "", Los_Angeles$weekly_price))## Warning: NAs introduced by coercionLos_Angeles$monthly_price_2 = as.numeric(gsub("\\$", "", Los_Angeles$monthly_price))## Warning: NAs introduced by coercionLos_Angeles_Pricing <- Los_Angeles %>% filter(price_2 <= 450)4 - Start plotting.
Here we have point plot that shows room types with prices as well as accomodates. The lower the accomodations, the more you see apartments. Once you get above 6, you see more houses and condominiums. You also see mostly homes priced at 400+.
ggplot(Los_Angeles, aes(x = price_2, y=accommodates, fill = property_type)) +
geom_point(size = 4, shape = 21, color = "white", stroke = 1.5) +
scale_fill_brewer(palette = "Paired") +
theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
labs(title="Rental types vs Pricing vs Accomdates", subtitle="Airbnb Los Angeles", caption="Source: Inside Airbnb",x="Price", y="Accomodates") + theme_bw() +
theme(panel.border = element_blank(), panel.grid.major = element_blank(), panel.grid.minor = element_blank(), axis.line = element_line(colour = "black"))## Warning: Removed 1 rows containing missing values (geom_point).
Here we have a scatterplot showing the number of reviews for hosts compared to the time when they became hosts. The colors separate hosts from superhosts. There doesn’t seem to be a trend with hosting and superhosts over time, but there are more superhosts than not have 150+ reviews.
ggplot(Los_Angeles, aes(x = host_since, y = number_of_reviews, color = host_is_superhost)) + geom_point()+ geom_abline() +
theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
labs(title="Number of Reviews Over Time", subtitle="Airbnb Los Angeles", caption="Source: Inside Airbnb",x="Host since", y="Number of reviews") + theme_bw() +
theme(panel.border = element_blank(), panel.grid.major = element_blank(), panel.grid.minor = element_blank(), axis.line = element_line(colour = "black"))
I am using this plot to touch upon chapter 5 of factfulness, the size instinct. There is a huge peak in this graph (that is not in my actual data so I don’t know how its showing up, but it works for this purpose) and it shows a host who charges close to 8000 for their Airbnb! One can see this image and think that perhaps more Airbnbs are like this. They can overexaggerate what the graph is displaying. The size of one individual and the visual associated seems so much bigger and more important than the rest, but its just one of over a hundred Airbnbs in the dataset.
As for the graph, it seems as though the people who have charged the most for Airbnbs are NOT superhosts, and many of them started using Airbnbing around mid2012 to mid2013.
ggplot(Los_Angeles_Pricing, aes(x = host_since, y = price_2, color = host_is_superhost)) + geom_col()+
theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
labs(title="Host vs Superhost Price Increases Over Time", subtitle="Airbnb Los Angeles", caption="Source: Inside Airbnb",x="Host since", y="Price") + theme_bw() +
theme(panel.border = element_blank(), panel.grid.major = element_blank(), panel.grid.minor = element_blank(), axis.line = element_line(colour = "black"))
Here we have a scatterplot showing prices on the y axis against cleaning fees on the x axis, and color representing rating reviews. The highest concentration of 100% reviews agglomerate around the $0-200 price range as well cleaning fees below $60. The concentration of 100% reviews spreads out the higher the price and cleaning fees become, and towards the more expensive end of the spectrum we start to see reviews in the 60% quantile. This might have to do with higher expectations from the clients for the higher prices, which if not met lead to greater disappointments.
We can apply the straight line instinct from factfulness to this figure. We see a positive trend for increased prices and cleaning fees leading to more lower percentaged views. Our instinct guides us into thinking that if we keep increasing prices and cleaning fees of airbnbs, then the review ratings will get lower and lower. However, this isn’t a fact, and if there are more expensive Airbnb listings that have higher cleaning rates, their client-base may be from a richer background, and the reviews would not necessarily be worse based on their experience. This is however, just another perspective, but it goes against my initial instinct, which is to think that increased prices means more disastisfied custumers.
ggplot() +
geom_point(data = Los_Angeles_Pricing, mapping = aes(x=cleaning_fee_2, y = price_2,
col = review_scores_rating ))+
scale_color_distiller(palette = "YlOrRd", direction = 1) +
labs(title="How Prices and Cleaning Fees Affect Reviews", subtitle="Airbnb Los Angeles", caption="Source: Inside Airbnb",x="Cleaning fee", y="Price") + theme_bw() +
theme(panel.border = element_blank(), panel.grid.major = element_blank(), panel.grid.minor = element_blank(), axis.line = element_line(colour = "black"))## Warning: Removed 10 rows containing missing values (geom_point).
Here I am attempting to use ggmap, so I had to create variables to create the fit for the boundaries of the map. This shows the latitude and longitude of Airbnbs in the data set, with a color legend showing price. There isn’t a particularly clear trend, but it does show cheaper Airbnbs are located away from the cost, in central LA/hollywood.
This is the first map that I’ve made in R, and I think it is just awesome!!
#ggmap - an object of class ggmap (from function get_map)
height <- max(Los_Angeles$latitude) - min(Los_Angeles$latitude)
width <- max(Los_Angeles$longitude) - min(Los_Angeles$longitude)
LA_borders <- c(bottom = min(Los_Angeles$latitude) - 0.1 * height,
top = max(Los_Angeles$latitude) + 0.1 * height,
left = min(Los_Angeles$longitude) - 0.1 * width,
right = max(Los_Angeles$longitude) + 0.1 * width)
map <- get_stamenmap(LA_borders, zoom = 10, maptype = "toner-lite")## Map from URL : http://tile.stamen.com/toner-lite/10/174/408.png## Map from URL : http://tile.stamen.com/toner-lite/10/175/408.png## Map from URL : http://tile.stamen.com/toner-lite/10/174/409.png## Map from URL : http://tile.stamen.com/toner-lite/10/175/409.pngggmap(map) +
geom_point(data = Los_Angeles, mapping = aes(x = longitude, y = latitude,
col = price_2)) +
scale_color_distiller(palette = "YlOrRd", direction = 1)