realestData: Population data - real estate in Poland at NUTS 4 level

In qape: Quantile of Absolute Prediction Errors

Population data - real estate in Poland at NUTS 4 level

Description

A data frame with 1504 observations on the following 7 variables (NUTS 4 units with masked values of the variables due to Statistical confidentiality has been removed).

Arguments

year	year.
NUTS4	NUTS 4 code (powiats).
NUTS2	NUTS 2 code (voivodships).
NUTS4type	type of NUTS 4 (1 - land counties, 2 - city counties/cities with powiat status).
premises	number of residential premises sold in market transactions (in thousands).
area	usable floor area of residential premises sold in market transactions (in millions of square meters).
price	sum of prices of residential premises sold (in billions of Polish zloty).

Source

Statitics Poland, https://bdl.stat.gov.pl/eng

Examples

data(realestData)
hist(realestData$price[realestData$year==2018])
boxplot(realestData$price~realestData$year)
boxplot(realestData$price[realestData$year==2018]~realestData$NUTS2[realestData$year==2018])
boxplot(realestData$price[realestData$year==2018]~realestData$NUTS4type[realestData$year==2018])


library(lme4)
attach(realestData)

N <- nrow(realestData[(year == 2015),]) # population size in the first period
n <- 75 # sample size in the first period
set.seed(123456)
sampled_elements_in_2015 <- sample(N,n) 
con2015 <- rep(0,N)
con2015[sampled_elements_in_2015] <- 1

sampled_elements_in_2016 <- sample(N,n)
con2016 <- rep(0,N)
con2016[sampled_elements_in_2016] <- 1

sampled_elements_in_2017 <- sample(N,n)
con2017 <- rep(0,N)
con2017[sampled_elements_in_2017] <- 1

sampled_elements_in_2018 <- sample(N,n)
con2018 <- rep(0,N)
con2018[sampled_elements_in_2018] <- 1

con=as.logical(con2015, con2016, con2017, con2018)

model1 <- lmer(price ~ premises + area + (1|NUTS2)+(0+premises|NUTS2) + 
(1|NUTS4type)+(0+area|NUTS4type), subset=con)
AIC(model1)
model2 <- lmer(price ~ premises + area + (0+premises|NUTS2) + (0+area|NUTS4type), subset = con)
AIC(model2)

qape documentation built on Aug. 21, 2023, 5:07 p.m.