-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathAirbnb_on_HousingValue.Rmd
296 lines (232 loc) · 9.54 KB
/
Airbnb_on_HousingValue.Rmd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
---
title: "Airbnb on Housing Values"
author: "Jonathan Eng"
output: pdf_document
date: " "
---
```{r, warning = FALSE, message = FALSE}
library(dplyr)
library(magrittr)
library(tidyverse)
library(plyr)
library(tibble)
library(stargazer)
```
```{r, warning = FALSE}
#import AirHomeData
AirHomeData2 = read.csv(file.path("cleaned_airbnb_home_data.csv"), header = TRUE)
```
```{r, WARNING = FALSE}
AirHomeData = AirHomeData2
#Remove all observations without the Y components "SALE.PRICE"
min_sale_price = 1
AirHomeData = AirHomeData[!is.na(AirHomeData$SALE.PRICE), ]
AirHomeData = AirHomeData[!(AirHomeData$SALE.PRICE < min_sale_price), ]
```
```{r}
#Remove Columns that are "Junk" or all NA'
AirHomeData %<>%
select(-c(ï..BOROUGH, BLOCK, LOT, EASE.MENT, ADDRESS, APARTMENT.NUMBER, ZIP.CODE, TOTAL.UNITS, BUILDING.CLASS.AT.PRESENT,
BUILDING.CLASS.AT.TIME.OF.SALE, COMMERCIAL.UNITS,TAX.CLASS.AT.TIME.OF.SALE, LAND.SQUARE.FEET,
Airbnb.Amount, HOME.AMOUNT,
))
AirHomeData %<>%
mutate(NEIGHBORHOOD = ifelse(NEIGHBORHOOD == 0, NA, NEIGHBORHOOD)) %<>%
mutate(BUILDING.CLASS.CATEGORY = ifelse(BUILDING.CLASS.CATEGORY == 0, NA, BUILDING.CLASS.CATEGORY)) %<>%
mutate(TAX.CLASS.AT.PRESENT = ifelse(TAX.CLASS.AT.PRESENT == 0, NA, TAX.CLASS.AT.PRESENT)) %<>%
mutate(YEAR.BUILT = ifelse(YEAR.BUILT == 0, NA, YEAR.BUILT)) %<>%
mutate(GROSS.SQUARE.FEET = ifelse(GROSS.SQUARE.FEET == 0, NA, GROSS.SQUARE.FEET)) %<>%
mutate(RESIDENTIAL.UNITS = ifelse(RESIDENTIAL.UNITS == 0, NA, RESIDENTIAL.UNITS)) %<>%
mutate(YEAR.BUILT = ifelse(YEAR.BUILT == 0, NA, YEAR.BUILT)) %<>%
mutate(PERCENT.AIRBNB = PERCENT.AIRBNB*100)
```
```{r}
AirHomeData %<>%
#Removes Extra Characters such as Dollar Signs ($) or Commas (,)
mutate(SALE.PRICE = gsub("[^0-9.-]", "", SALE.PRICE)) %<>%
mutate(YEAR.BUILT = gsub("[^0-9.-]", "", YEAR.BUILT)) %<>%
mutate(RESIDENTIAL.UNITS = gsub("[^0-9.-]", "", RESIDENTIAL.UNITS)) %<>%
mutate(GROSS.SQUARE.FEET = gsub("[^0-9.-]", "", GROSS.SQUARE.FEET)) %<>%
#Log Sale Price and GrossSquareFeet
mutate(SALE.PRICE = log(as.double(SALE.PRICE))) %<>%
mutate(GROSS.SQUARE.FEET = log(as.integer(GROSS.SQUARE.FEET))) %<>%
mutate(YEAR.BUILT = as.factor(YEAR.BUILT)) %<>%
mutate(RESIDENTIAL.UNITS = as.integer(RESIDENTIAL.UNITS)) %<>%
mutate(NEIGHBORHOOD = as.factor(NEIGHBORHOOD)) %<>%
mutate(BUILDING.CLASS.CATEGORY = as.factor(BUILDING.CLASS.CATEGORY)) %<>%
mutate(TAX.CLASS.AT.PRESENT = as.factor(TAX.CLASS.AT.PRESENT)) %<>%
mutate(SALE.DATE = as.integer(substr(SALE.DATE, nchar(SALE.DATE)-3, nchar(SALE.DATE)))) %<>%
dplyr::rename(SALE.YEAR = SALE.DATE) %<>%
dplyr::rename(LOG_GROSS.SQUARE.FEET = GROSS.SQUARE.FEET) %<>%
dplyr::rename(LOG_SALE.PRICE = SALE.PRICE) %<>%
dplyr::rename(TAX_CLASS = TAX.CLASS.AT.PRESENT) %<>%
mutate(SALE.YEAR = as.factor(SALE.YEAR))
```
```{r}
AirHomeData$NEIGHBORHOOD = droplevels(AirHomeData$NEIGHBORHOOD)
AirHomeData$BUILDING.CLASS.CATEGORY = droplevels(AirHomeData$BUILDING.CLASS.CATEGORY)
AirHomeData$TAX_CLASS = droplevels(AirHomeData$TAX_CLASS)
AirHomeData$YEAR.BUILT = droplevels(AirHomeData$YEAR.BUILT)
AirHomeData$SALE.YEAR = droplevels(AirHomeData$SALE.YEAR)
```
```{r}
#Removes Outliers using IQR Method
AirHomeData = na.omit(AirHomeData)
temp2 = AirHomeData
Q1 = quantile(AirHomeData$LOG_SALE.PRICE, prob = .25)
Q3 = quantile(AirHomeData$LOG_SALE.PRICE, prob = .75)
LowerBound = Q1 - 1.5*IQR(AirHomeData$LOG_SALE.PRICE)
UpperBound = Q3 + 1.5*IQR(AirHomeData$LOG_SALE.PRICE)
AirHomeData = AirHomeData[AirHomeData$LOG_SALE.PRICE >= LowerBound, ]
AirHomeData = AirHomeData[AirHomeData$LOG_SALE.PRICE <= UpperBound, ]
```
```{r}
#MIN MAX MEAN Summaries
cat("IQR METHOD: \n\n")
cat("Minimum Values: \n")
apply(AirHomeData,2,min)
cat("Maximum Values: \n")
apply(AirHomeData,2,max)
cat("\nAverage RESIDENTIAL.UNITS: ", mean(AirHomeData$RESIDENTIAL.UNITS))
cat("\nAverage LOG_GROSS.SQUARE.FEET: ", mean(AirHomeData$LOG_GROSS.SQUARE.FEET))
cat("\nAverage Sale.Price: ", mean(AirHomeData$LOG_SALE.PRICE))
cat("\nAverage Percent.Airbnb: ", mean(AirHomeData$PERCENT.AIRBNB))
```
```{r}
# Clean Factors without *n* many Values
n = 500
AirHomeData %<>%
group_by(BUILDING.CLASS.CATEGORY) %<>%
filter(n() >= n)
AirHomeData %<>%
group_by(NEIGHBORHOOD) %<>%
filter(n() >= n)
AirHomeData %<>%
group_by(TAX_CLASS) %<>%
filter(n() >= n)
AirHomeData %<>%
group_by(YEAR.BUILT) %<>%
filter(n() >= n)
```
```{r}
AirHomeData$NEIGHBORHOOD = droplevels(AirHomeData$NEIGHBORHOOD)
AirHomeData$BUILDING.CLASS.CATEGORY = droplevels(AirHomeData$BUILDING.CLASS.CATEGORY)
AirHomeData$TAX_CLASS = droplevels(AirHomeData$TAX_CLASS)
AirHomeData$YEAR.BUILT = droplevels(AirHomeData$YEAR.BUILT)
AirHomeData$SALE.YEAR = droplevels(AirHomeData$SALE.YEAR)
```
```{r, WARNING = FALSE}
AirHomeDataTemp = model.matrix( ~ ., data=AirHomeData, contrasts.arg =
lapply(data.frame(AirHomeData[,sapply(data.frame(AirHomeData), is.factor)]), contrasts, contrasts = FALSE))
AirHomeDataTemp2 = data.frame(AirHomeDataTemp)
#INTERCEPTS:
AirHomeDataTemp2 %<>%
select(-c( X.Intercept.,
NEIGHBORHOODWOODHAVEN,
BUILDING.CLASS.CATEGORY07..RENTALS...WALKUP.APARTMENTS,
YEAR.BUILT1899,
TAX_CLASS2B,
SALE.YEAR2013,
))
```
```{r}
#Y = Response Variable; X = Features
y = AirHomeDataTemp2$LOG_SALE.PRICE
X = AirHomeDataTemp2 %>%
select(-c(LOG_SALE.PRICE))
```
```{r}
#Set New Data
newdata = cbind(X, y)
newdata %<>%
dplyr::rename(LOG_SALE.PRICE = y) %<>%
select(LOG_SALE.PRICE,
everything())
X = newdata[ ,2:ncol(newdata)]
y = newdata[ ,1]
```
```{r, warning = FALSE}
n = nrow(newdata)
K = n/100
test_indices = sample(1 : n, size = n * 1 / K)
master_train_indices = setdiff(1 : n, test_indices)
select_indices = sample(master_train_indices, size = n * 1 / K)
train_indices = setdiff(master_train_indices, select_indices)
rm(master_train_indices)
pacman::p_load(testthat)
expect_equal(1 : n, sort(c(train_indices, select_indices, test_indices)))
newdata_train = newdata[train_indices, ]
newdata_select = newdata[select_indices, ]
newdata_test = newdata[test_indices, ]
rm(test_indices, select_indices, train_indices)
mod1 = lm(LOG_SALE.PRICE ~ ., newdata_train)
mod2 = lm(LOG_SALE.PRICE ~ poly(PERCENT.AIRBNB, 2) + poly(LOG_GROSS.SQUARE.FEET, 2) + poly(RESIDENTIAL.UNITS,2 ) + ., newdata_train)
mod3 = lm(LOG_SALE.PRICE ~ poly(PERCENT.AIRBNB, 3) + poly(LOG_GROSS.SQUARE.FEET, 3) + poly(RESIDENTIAL.UNITS,3 ) + ., newdata_train)
y_select = newdata$LOG_SALE.PRICE #True Price
yhat_select_mod1 = predict(mod1, newdata_select)
yhat_select_mod2 = predict(mod2, newdata_select)
yhat_select_mod3 = predict(mod3, newdata_select)
s_e_s = c(sd(yhat_select_mod1 - y_select), sd(yhat_select_mod2 - y_select), sd(yhat_select_mod3 - y_select))
names(s_e_s) = paste("mod", 1:3, sep = "")
s_e_s
names(which.min(s_e_s))
yhat_test_mod1 = predict(mod1, newdata_test)
y_test = newdata_test$LOG_SALE.PRICE #True Price
sd(yhat_test_mod1 - y_test)
mod_final = lm(LOG_SALE.PRICE ~ . , newdata)
summary(mod_final)
rss = c(crossprod(mod_final$residuals))
mse = rss / length(mod_final$residuals)
rmse_kfolds = sqrt(mse)
cat("\nLinear Regression RMSE: ", rmse_kfolds)
```
```{r, warning = FALSE}
#Values in the Intercept:
#NEIGHBORHOOD.ARDEN.HEIGHTS
#BUILDING.CLASS.CATEGORY01.ONE.FAMILY.DWELLINGS
#YEAR.BUILT1899,
#TAX.CLASS.AT.PRESENT1,
#SALE.YEAR2013,
nvar = 2
model1 = lm(LOG_SALE.PRICE ~ PERCENT.AIRBNB, data = AirHomeData) #Short Term Rental Concentration
model1$coefficients = model1$coefficients[2:nvar]
nvar = 4
model2 = lm(LOG_SALE.PRICE ~ PERCENT.AIRBNB + #Short Term Rental Concentration
LOG_GROSS.SQUARE.FEET + RESIDENTIAL.UNITS, #Building/Home Size/Units
data = AirHomeData)
model2$coefficients = model2$coefficients[2:nvar]
#Multilevel Factors Introduced
nvar = 10
model3 = lm(LOG_SALE.PRICE ~ PERCENT.AIRBNB + #Short Term Rental Concentration
LOG_GROSS.SQUARE.FEET + RESIDENTIAL.UNITS + #Building/Home Size/Units
TAX_CLASS + BUILDING.CLASS.CATEGORY, #Building/Home Features
data = AirHomeData)
model3$coefficients = model3$coefficients[2:nvar]
nvar = 30
model4 = lm(LOG_SALE.PRICE ~ PERCENT.AIRBNB + #Short Term Rental Concentration
LOG_GROSS.SQUARE.FEET + RESIDENTIAL.UNITS + #Building/Home Size/Units
TAX_CLASS + BUILDING.CLASS.CATEGORY + #Building/Home Features
YEAR.BUILT, #Year Built
data = AirHomeData)
model4$coefficients = model4$coefficients[2:nvar]
nvar = 32
model5 = lm(LOG_SALE.PRICE ~ PERCENT.AIRBNB + #Short Term REntal Concentration
LOG_GROSS.SQUARE.FEET + RESIDENTIAL.UNITS + #Building/Home Size/Units
TAX_CLASS + BUILDING.CLASS.CATEGORY + #Building/Home Features
YEAR.BUILT + #Year Built
SALE.YEAR, #Sale Year
data = AirHomeData)
model5$coefficients = model5$coefficients[2:nvar]
nvar = 100
model6 = lm(LOG_SALE.PRICE ~ PERCENT.AIRBNB + #Short Term Rental Concentration
LOG_GROSS.SQUARE.FEET + RESIDENTIAL.UNITS + #Building/Home Size/Units
TAX_CLASS + BUILDING.CLASS.CATEGORY + #Building/Home Features
YEAR.BUILT + #Year Built
SALE.YEAR + #Sale Year
NEIGHBORHOOD, #Neighborhood Fixed Effects
data = AirHomeData)
model6$coefficients = model6$coefficients[2:nvar]
```
```{r, warning = FALSE}
stargazer(model1, model2, model3, model4, model5, model6, title = "Home Sale Price versus Short Term Rental Concentration", type = "text", out = "Home Sale Price versus Short Term Rental Concentration")
```