#Number of apartments and price in Belgrade
library(readxl)
df <- read_xlsx("C:/Users/vjovanovic/Desktop/R Udemy/Vezbe sa javnim podacima/stanovi.xlsx")
str(df)
## Classes 'tbl_df', 'tbl' and 'data.frame': 3270 obs. of 8 variables:
## $ idindikator: chr "050502IND01" "050502IND01" "050502IND01" "050502IND01" ...
## $ mes : chr "00" "00" "00" "00" ...
## $ god : chr "2011" "2011" "2018" "2018" ...
## $ IDVrPod : chr "1" "2" "1" "2" ...
## $ nVrPod : chr "Broj završenih stanova - ukupno" "Površina završenih stanova - ukupno, m2" "Broj završenih stanova - ukupno" "Površina završenih stanova - ukupno, m2" ...
## $ IDTer : chr "70017" "70017" "70017" "70017" ...
## $ nTer : chr "Aleksandrovac" "Aleksandrovac" "Aleksandrovac" "Aleksandrovac" ...
## $ vrednost : chr "7" "883" "16" "1671" ...
#Now we must change character to factor or numeric variable
df$nTer <- as.factor(df$nTer)
df$vrednost <- as.numeric(df$vrednost)
df$god <- as.factor(df$god)
df$IDVrPod <- as.factor(df$IDVrPod)
head(df)
## # A tibble: 6 x 8
## idindikator mes god IDVrPod nVrPod IDTer nTer vrednost
## <chr> <chr> <fct> <fct> <chr> <chr> <fct> <dbl>
## 1 050502IND01 00 2011 1 Broj završenih st~ 70017 Aleksa~ 7
## 2 050502IND01 00 2011 2 Površina završeni~ 70017 Aleksa~ 883
## 3 050502IND01 00 2018 1 Broj završenih st~ 70017 Aleksa~ 16
## 4 050502IND01 00 2018 2 Površina završeni~ 70017 Aleksa~ 1671
## 5 050502IND01 00 2017 1 Broj završenih st~ 70017 Aleksa~ 6
## 6 050502IND01 00 2017 2 Površina završeni~ 70017 Aleksa~ 982
#Now we will select only variables for the Republic of Serbia, not for particular cities.
library(dplyr)
##
## Attaching package: 'dplyr'
## The following objects are masked from 'package:stats':
##
## filter, lag
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union
df1 <- filter(df, nTer=="Beogradski region")
head(df1)
## # A tibble: 6 x 8
## idindikator mes god IDVrPod nVrPod IDTer nTer vrednost
## <chr> <chr> <fct> <fct> <chr> <chr> <fct> <dbl>
## 1 050502IND01 00 2011 1 Broj završenih s~ RS11 Beograd~ 6416
## 2 050502IND01 00 2011 2 Površina završen~ RS11 Beograd~ 379681
## 3 050502IND01 00 2012 1 Broj završenih s~ RS11 Beograd~ 8096
## 4 050502IND01 00 2012 2 Površina završen~ RS11 Beograd~ 486739
## 5 050502IND01 00 2013 1 Broj završenih s~ RS11 Beograd~ 7596
## 6 050502IND01 00 2013 2 Površina završen~ RS11 Beograd~ 484348
#We must cast the variable vrednost
library(reshape)
##
## Attaching package: 'reshape'
## The following object is masked from 'package:dplyr':
##
## rename
recasteddf <- cast(df1, idindikator + god~nVrPod,sum)
## Using vrednost as value column. Use the value argument to cast to override this choice
head(recasteddf)
## idindikator god Broj završenih stanova - ukupno
## 1 050502IND01 2011 6416
## 2 050502IND01 2012 8096
## 3 050502IND01 2013 7596
## 4 050502IND01 2014 4755
## 5 050502IND01 2015 4014
## 6 050502IND01 2016 3167
## Površina završenih stanova - ukupno, m2
## 1 379681
## 2 486739
## 3 484348
## 4 301457
## 5 245786
## 6 223856
str(recasteddf)
## List of 4
## $ idindikator : chr [1:8] "050502IND01" "050502IND01" "050502IND01" "050502IND01" ...
## $ god : Factor w/ 8 levels "2011","2012",..: 1 2 3 4 5 6 7 8
## $ Broj završenih stanova - ukupno : num [1:8] 6416 8096 7596 4755 4014 ...
## $ Površina završenih stanova - ukupno, m2: num [1:8] 379681 486739 484348 301457 245786 ...
## - attr(*, "row.names")= int [1:8] 1 2 3 4 5 6 7 8
## - attr(*, "idvars")= chr [1:2] "idindikator" "god"
## - attr(*, "rdimnames")=List of 2
## ..$ :'data.frame': 8 obs. of 2 variables:
## .. ..$ idindikator: chr [1:8] "050502IND01" "050502IND01" "050502IND01" "050502IND01" ...
## .. ..$ god : Factor w/ 8 levels "2011","2012",..: 1 2 3 4 5 6 7 8
## ..$ :'data.frame': 2 obs. of 1 variable:
## .. ..$ nVrPod: chr [1:2] "Broj završenih stanova - ukupno" "Površina završenih stanova - ukupno, m2"
#Now we will plot a graph
library(ggplot2)
p <- ggplot(recasteddf, aes(god, `Broj završenih stanova - ukupno`)) + geom_point(alpha=0.5, size=3) + theme_bw()
p
g<-p+geom_col(alpha=0.5, size=4, fill="blue") + xlab("Year") + ylab("Number of new appartments in Belgrade") + ggtitle("The Growth of appartments in last 30 years in Belgrade")
print(g)
j <- ggplot(recasteddf, aes(god, `Površina završenih stanova - ukupno, m²`)) + geom_point(alpha=0.5, size=3) + theme_bw()
j
k<-j+geom_col(alpha=0.5, size=4, fill="blue") + xlab("Year") + ylab("Area of new appartments in Belgrade") + ggtitle("The Area of new appartments in last 30 years in Belgrade")
k
