1 Preface
2 Import datasets
3 Internal validation
4 External validation
- 4.1 Data Completness: Number of buildings by urban district
- 4.2 Data plausibility: Number of buildings by floor counts

1 Preface

This code corresponds with the Technical Validation section in the Data Descriptor “Building schematic of Vienna in the late 1920s”, published by Nature Scientific Data.

Please consider the following steps to run the code 1. Create a new directory on your computer (e.g. “c:/building.schematic”) 2. Download the files from the Github repository and save them in your new directory. 3. Copy the path of your new directory into the code at line 42.

2 Import datasets

######################################################################################
path <- "C:/Users/u.kral/ownCloud/03_TU Wien/Github/building.schematic/"
######################################################################################


# Import file "Dataset.csv", which is the digital building schematic.
dataset <- read.csv(file=paste(path, "Dataset.csv", sep = ""), sep = ";", stringsAsFactors = FALSE, encoding = "UTF-8")

# Import file "Online-Only Table 2.csv", which is identical with Online-only Table 2 in the Data Descriptor.
cadastral_raw <- read.csv(file=paste(path, "/Data.descriptor/Online-only Table 2.csv", sep=""),  sep = ";" , stringsAsFactors = F)
cadastral <- cadastral_raw[1:66,] # Cadastral communities mentioned in the analog building schematic

# Import file "adressen_standorte_wien_20201015.csv". This file includes today's street names in the city of Vienna. [Open data Österreich](https://www.data.gv.at/katalog/dataset/stadt-wien_adressdatenderstadtwien)
adressen <- read.csv(file=paste(path, "/Data.descriptor/adressen_standorte_wien_20201015.csv", sep=""), sep = ";", stringsAsFactors = F,fileEncoding = "UTF-8")

# Import file "statistical_yearbook (1914).xlsx". Data retrieved from digitized report [Statistisches Jahrbuch der Stadt Wien. Bd. 1914](https://www.digital.wienbibliothek.at/wbrobv/periodical/titleinfo/2057276)
floors_1914 <- read_xlsx(paste(path, "/Data.descriptor/statistical_yearbook (1914).xlsx", sep=""), sheet = "STKW_hist", col_names = TRUE, range = "B7:I27")

colnames(floors_1914) <- c("UD.1920s", "FLOORS_0", "FLOORS_1", "FLOORS_2", "FLOORS_3", "FLOORS_4", "FLOORS_5", "FLOORS_unknown")

# Import file "statistical_yearbook (1923).xlsx". Data retrieved from digitized report [Statistisches Jahrbuch der Stadt Wien. Bd. 1929 (1. Jahrgang)](https://www.digital.wienbibliothek.at/wbrobv/periodical/titleinfo/2057276)
yearbook_1923 <- read_xlsx(paste(path, "/Data.descriptor/statistical_yearbook (1923).xlsx", sep=""), sheet = "Rohdaten", col_names =  TRUE, col_types = rep("numeric", times = 2))

3 Internal validation

This code section produces Figure 6 in the Data Descriptor.

3.1 ID

# Create categories
id_0 <- grep("_[0-9]*", dataset$ID) # unfolded IDs
id_1 <- grep("^[0-9]*$", dataset$ID) # non-unfolded IDs

id_plot_data <- dataset
id_plot_data$id_flag <- rep(NA, nrow(dataset))
id_plot_data[id_0, "id_flag"] <- "yes" # logic: ID unfolded: yes
id_plot_data[id_1, "id_flag"] <- "no"

# Generate plot
id_plot_data_grouped <- id_plot_data %>%
  group_by(UD.1920s, id_flag) %>%
  summarize(count = n())

id_plot_data_grouped <- data.frame(id_plot_data_grouped, stringsAsFactors = F)

totals<- id_plot_data_grouped %>%
    group_by(UD.1920s) %>%
    summarise(total = sum(count))

id_plot <- ggplot(id_plot_data_grouped, aes(x=UD.1920s, y=count, fill= factor(id_flag, levels = c("yes", "no")))) +
  geom_bar(stat="identity", position="stack")+
  labs(title = "ID", x="Urban district (UD.1920s)", y="Number of data entries")+
  theme(plot.title = element_text(face = "bold"))+
  scale_fill_manual(name = "ID unfolded", values=c('lightblue','darkblue'))+
  geom_text(data=totals, aes(x=UD.1920s, label=total, y=total, fill=NULL), nudge_y=150, size = 3)+
  scale_x_continuous(breaks = c(1:21))
id_plot

3.2 STR.1920s

str_ma37 <- as.character(levels(as.factor(adressen$NAME_STR)))

# Standardizing the street names in "Adressen Standorte Wien", which is the external datasource for validating STR.1920s and STR.2010s names in the dataset
str_m <- gsub(" ","", str_ma37) 
str_m <- lowerCase(str_m) 
str_m <- gsub("st\\.", "sankt", str_m)
str_m <- gsub("-","",str_m) 
str_m <- gsub("dr\\.","dr",str_m) 
str_m <- gsub("ß","ss",str_m)
str_m <- gsub("'","",str_m)
str_m <- gsub("\\.","",str_m)

str_gesamt <- data.frame(str_m, str_m, stringsAsFactors = FALSE)
colnames(str_gesamt) <- c("str_gesamt", "str_ma37")

# Standardizing the STR.1920s names in the dataset
dataset$str_1920_norm <- dataset$STR.1920s
dataset$str_1920_norm <- lowerCase(dataset$str_1920_norm) 
dataset$str_1920_norm <- gsub("ß","ss", dataset$str_1920_norm)
dataset$str_1920_norm <- gsub("st\\.","sankt",dataset$str_1920_norm)
dataset$str_1920_norm <- gsub("dr\\.","dr",dataset$str_1920_norm)
dataset$str_1920_norm <- gsub(" ","", dataset$str_1920_norm) 
dataset$str_1920_norm <- gsub("'","",dataset$str_1920_norm)
dataset$str_1920_norm <- gsub("\\.","", dataset$str_1920_norm)
dataset$str_1920_norm <- gsub("-","", dataset$str_1920_norm)
dataset$str_1920_norm <- gsub("\\(","", dataset$str_1920_norm)
dataset$str_1920_norm <- gsub("\\)","", dataset$str_1920_norm)

dataset$str_2010_norm <- dataset$STR.2010s
dataset$str_2010_norm <- lowerCase(dataset$str_2010_norm) # alles auf Kleinbuchstaben
dataset$str_2010_norm <- gsub("ß","ss", dataset$str_2010_norm)
dataset$str_2010_norm <- gsub("st\\.","sankt",dataset$str_2010_norm)
dataset$str_2010_norm <- gsub("dr\\.","dr",dataset$str_2010_norm)
dataset$str_2010_norm <- gsub(" ","", dataset$str_2010_norm) # Leerzeichen entfernen
dataset$str_2010_norm <- gsub("'","",dataset$str_2010_norm)
dataset$str_2010_norm <- gsub("\\.","", dataset$str_2010_norm)
dataset$str_2010_norm <- gsub("-","", dataset$str_2010_norm)
dataset$str_2010_norm <- gsub("\\(","", dataset$str_2010_norm)
dataset$str_2010_norm <- gsub("\\)","", dataset$str_2010_norm)

# Assigning standardized names from "Adressen Standorte Wien" to the standardized names of STR.2010s.
dataset <- merge(dataset, str_gesamt, by.x = "str_2010_norm", by.y = "str_gesamt", all.x = TRUE)

# Assigning standardized names from "Adressen Standorte Wien" to the standardized names of STR.1920s.
dataset <- merge(dataset, str_gesamt, by.x = "str_1920_norm", by.y = "str_gesamt", all.x = TRUE)

check_length_ma37.y <- length(levels(as.factor(dataset$str_ma37.y))) 

match_yes<- dataset[which(dataset$str_ma37.y != ""),] 
match_no <- dataset[which(is.na(dataset$str_ma37.y) == T),] 

match_yes_control <- dataset[which(dataset$str_1920_norm == dataset$str_2010_norm), ] 
match_no_control <- dataset[which(dataset$str_1920_norm != dataset$str_2010_norm), ] 

test1 <- match_no_control[which(is.na(match_no_control$str_ma37.y) == F & is.na(match_no_control$str_ma37.x) == F),] # Street name has been changed.
test2 <- match_no_control[which(is.na(match_no_control$str_ma37.y) == F & is.na(match_no_control$str_ma37.x) == T),] # No STR.2010s counterpart or removed from the street name register.
test3 <- match_no_control[which(is.na(match_no_control$str_ma37.y) == T &is.na(match_no_control$str_ma37.x) == F),] # Street name has been changed.
test4 <- match_no_control[which(is.na(match_no_control$str_ma37.y) == T &is.na(match_no_control$str_ma37.x) == T),] # No STR.2010s counterpart or removed from the street name register.

# unique records
test1_uni <- unique(test1[,c("str_1920_norm","str_2010_norm")])
test2_uni <- unique(test2[,c("str_1920_norm","str_2010_norm")])
test3_uni <- unique(test3[,c("str_1920_norm","str_2010_norm")])
test4_uni <- unique(test4[,c("str_1920_norm","str_2010_norm")])

# Numbers for the data descriptor
str_1920_spelling_rows <- cbind(c(nrow(match_yes_control), nrow(match_no_control ), nrow(test1), nrow(test2), nrow(test3), nrow(test4)), round(c(nrow(match_yes_control), nrow(match_no_control ), nrow(test1), nrow(test2), nrow(test3), nrow(test4))/nrow(dataset)*100,2))

dataset$str_1920_spelling = c(rep(NA, nrow(dataset)))

# Data preparation for barplots
pos_str_1920_spelling_1 <- c(which(dataset$str_1920_norm == dataset$str_2010_norm), which(is.na(match_no_control$str_ma37.y) == F & is.na(match_no_control$str_ma37.x) == F), which(is.na(match_no_control$str_ma37.y) == F & is.na(match_no_control$str_ma37.x) == T))

pos_str_1920_spelling_2 <- c(which(is.na(match_no_control$str_ma37.y) == T &is.na(match_no_control$str_ma37.x) == F),which(is.na(match_no_control$str_ma37.y) == T &is.na(match_no_control$str_ma37.x) == T))

control_length_str_1920 <- length(pos_str_1920_spelling_1)+length(pos_str_1920_spelling_2)-nrow(dataset)

str_1920_spelling_t1 <- c(match_yes_control[,"ID"], test1[,"ID"], test2[,"ID"])
str_1920_spelling_t2 <- c(test3[,"ID"], test4[,"ID"])


dataset[(dataset$ID %in% str_1920_spelling_t1), "str_1920_spelling"] <- "1" # Adressen Standorte Wien
dataset[(dataset$ID %in% str_1920_spelling_t2), "str_1920_spelling"] <- "2" # # Wien Geschichte Wiki, Analog Building Schematic

rows_pos_str_1920_spelling <- dataset %>%
  group_by(UD.1920s, str_1920_spelling) %>%
  summarize(count_spelling_1920 = n())

# Generate plot
dataplot <- data.frame(rows_pos_str_1920_spelling, stringsAsFactors = F)
dataplot <- rbind(dataplot, c(8,2,0))
dataplot <- dataplot[order(dataplot$UD.1920s, dataplot$str_1920_spelling),]

totals<- dataplot %>%
    group_by(UD.1920s) %>%
    summarise(total=sum(count_spelling_1920))

str_1920_plot <- ggplot(dataplot, aes(x=UD.1920s, y=count_spelling_1920, fill= factor(str_1920_spelling, levels = c("2","1")))) +
  geom_bar(stat="identity", position="stack") +
  labs(title = "STR.1920s", x="Urban district (UD.1920s)", y="Number of data entries") +
  scale_fill_manual(name = "Name spelling\nverified by", labels = c("Digital building schematic,\nWien Geschichte Wiki", "Adressen Standorte Wien"), values=c('lightblue','darkblue')) +
  geom_text(data=totals, aes(x=UD.1920s, label=total, y=total, fill=NULL), nudge_y=150, size = 3) +
  scale_x_continuous(breaks = c(1:21)) +
  theme(plot.title = element_text(face = "bold"))
str_1920_plot

count_spelling_1920_sum <- nrow(dataset) - sum(dataplot$count_spelling_1920)

3.3 STR.2010s

# Create categories
str_2010_1 <- dataset[which(is.na(dataset$str_ma37.x) == FALSE),] 
str_2010_2 <- dataset[which(is.na(dataset$str_ma37.x) == TRUE),]

str_2010_sum <- nrow(dataset) - length(str_2010_1) - length(str_2010_2)

str_2010_spelling_t1 <- str_2010_1[,"ID"]
str_2010_spelling_t2 <- str_2010_2[,"ID"]

dataset[(dataset$ID %in% str_2010_spelling_t1), "str_2010_spelling"] <- "1" # Adressen Standorte Wien
dataset[(dataset$ID %in% str_2010_spelling_t2), "str_2010_spelling"] <- "2" # no STR.2010s

rows_pos_str_2010_spelling <- dataset %>%
  group_by(UD.1920s, str_2010_spelling) %>%
  summarize(count_spelling_2010 = n())

# Generate plot
dataplot <- data.frame(rows_pos_str_2010_spelling, stringsAsFactors = F)
dataplot <- rbind(dataplot, c(8,2,0))
dataplot <- dataplot[order(dataplot$UD.1920s, dataplot$str_2010_spelling),]

totals<- dataplot %>%
    group_by(UD.1920s) %>%
    summarise(total=sum(count_spelling_2010))


# Plot
str_2010_plot <- ggplot(dataplot, aes(x=UD.1920s, y=count_spelling_2010, fill= factor(str_2010_spelling, levels = c("2","1")))) +
  geom_bar(stat="identity", position="stack") +
  labs(title = "STR.2010s", x="Urban district (UD.1920s)", y="Number of data entries") +
  scale_fill_manual(name = "Name spelling\nverified by", labels = c("not relevant, because no STR.2010s\ncounterpart from STR.1920s", "Adressen Standorte Wien"), values=c('#999999','darkblue')) +
  geom_text(data=totals, aes(x=UD.1920s, label=total, y=total, fill=NULL), nudge_y=150, size = 3) +
  scale_x_continuous(breaks = c(1:21))+
  theme(plot.title = element_text(face = "bold"))
str_2010_plot

count_spelling_2010_sum <- nrow(dataset) - sum(dataplot$count_spelling_2010)

3.4 UD.1920

ud <- dataset %>%
  group_by(UD.1920s) %>%
  summarize(count = n())
ud <- data.frame(ud, stringsAsFactors = F)

ud_merge <- merge(ud, unique(cadastral[,c("UD.1920s", "Volume")]), by.x = "UD.1920s", by.y = "UD.1920s", sort = F)
ud_merge <- data.frame(ud_merge, stringsAsFactors = F)

ud_plot <-ggplot(data=ud_merge, aes(x=UD.1920s, y=count, fill = factor(Volume, levels = c(1:10)))) +
  geom_bar(stat="identity", width=.8) +
  geom_text(aes(label=count), nudge_y=120, size = 3)+
  labs(title = "UD.1920s", x="Urban district (UD.1920s)", y="Number of data entries") +
  scale_x_continuous(breaks = c(1:21))+
  theme(plot.title = element_text(face = "bold"))+
  scale_fill_discrete("Volume of analog\nbuilding schematic")
ud_plot

ud_sum <- nrow(dataset) - sum(ud$count)

3.5 CC.2010s

# Create categories
cc <- dataset %>%
  group_by(CC.2010s) %>%
  summarize(count = n())
cc <- data.frame(cc, stringsAsFactors = F)
cc$CC.2010s <- as.character(cc$CC.2010s)

cadastral_raw_sub <- unique(cadastral_raw[,c("Volume", "cadastral.number_2010s")])
cadastral_raw_sub$cadastral.number_2010s <- as.character(cadastral_raw_sub$cadastral.number_2010s)

cc_merge <- merge(cc, cadastral_raw_sub, by.x = "CC.2010s", by.y = "cadastral.number_2010s", by=all, sort = F)

# Generate plot

totals<- cc  %>%
    group_by(CC.2010s) %>%
    summarise(total=sum(count))

cc_plot <-ggplot(data=cc_merge, aes(x=CC.2010s, y=count, fill = factor(Volume, levels = c(1:10)))) +
  geom_bar(stat="identity") +
  geom_text(aes(label=count), vjust=0.3, hjust=-.5, size=2.5, angle = 90) +
  labs(title = "CC.2010s", x="Cadastral communites (CC.2010s)",y="Number of data entries") +
  theme(axis.text.x=element_text(angle = 90, size = 6),plot.title = element_text(face = "bold"))+
  coord_cartesian(ylim = c(0,round(max(cc$count),-3)))+
  scale_fill_discrete(name = "Volume of analog\nbuilding schematic")
cc_plot

cc_sum <- nrow(dataset) - sum(cc_merge$count)

3.6 BN.1920s

# Create categories

pos1 <- grep("\\D$", dataset$BN.1920s) # 
test1 <- levels(as.factor(dataset[pos1, "BN.1920s"])) # 118b

pos2 <- grep("\\d$", dataset$BN.1920s) # 
test2 <- levels(as.factor(dataset[pos2, "BN.1920s"])) # 118 inkl. 2 Einträge für "neben27"

pos3 <- which(dataset$BN.1920s =="")
test3 <- dataset[pos3,]

pos4 <- grep("^[A-Za-z]", dataset$BN.1920s)
test4 <- levels(as.factor(dataset[pos4, "BN.1920s"])) # 2 Einträge für "neben27"

# die "neben27" rausfiltern
t <- which(pos2 %in% pos4)
pos2 <- pos2[-t] 

pos_comb <- c(pos1, pos2, pos3)
pos_c <- length(pos1)+length(pos2)+length(pos3)

bn.count <- c(length(pos2)-2, length(pos1), length(pos4), length(pos3))
bn.discr <- c("Only Integer", "Integer and letters", "Letters and integers", "Data not available")

bn.table <- cbind(bn.discr, bn.count)
bn.table <- as.data.frame(bn.table, stringsAsFactors = F)
bn.table$bn.count <- as.numeric(bn.table$bn.count)
bn.table$rel <- bn.table$bn.count / sum(bn.table$bn.count)*100

dataset$bn.table.ud = rep(NA, nrow(dataset))
dataset[pos1,"bn.table.ud"] <- bn.discr[2]
dataset[pos2,"bn.table.ud"] <- bn.discr[1]
dataset[pos3,"bn.table.ud"] <- bn.discr[4]
dataset[pos4,"bn.table.ud"] <- bn.discr[3]

bn.table_ud <- dataset %>%
  group_by(UD.1920s, bn.table.ud) %>%
  summarize(count = n())

test_sum <- nrow(dataset)- sum(bn.table_ud$count)

# Generate plot
bn.table_ud <- data.frame(bn.table_ud, stringsAsFactors = F)

totals<- bn.table_ud %>%
    group_by(UD.1920s) %>%
    summarize(total=sum(count))

position_bn <- levels(as.factor(bn.table_ud$bn.table.ud))

bn_plot <-ggplot(data=bn.table_ud, aes(x=UD.1920s, y=count, fill = factor(bn.table.ud, levels = position_bn[c(1,3,2,4)]))) +
  geom_bar(stat="identity") +
  scale_fill_manual(name = "Data pattern", values=c('#999999','#87CEFA','#4169E1','darkblue')) +
  labs(title = "BN.1920s", x="Urban district (UD.1920s)", y="Number of data entries") +
  geom_text(data=totals, aes(x=UD.1920s, label=total, y=total, fill=NULL), nudge_y=150, size = 3)+
  scale_x_continuous(breaks = c(1:21))+
  theme(plot.title = element_text(face = "bold"))
bn_plot

3.7 AREA.1920s

# Create categories
area_pos_yes <- which(dataset$AREA.1920s != "")
area_pos_no <- which(is.na(dataset$AREA.1920s) == TRUE)

dataset$area_flag <- c(rep(NA, nrow(dataset)))
dataset[area_pos_yes, "area_flag"] <- "yes"
dataset[area_pos_no, "area_flag"] <- "no"

area_plot3 <- dataset %>%
  group_by(UD.1920s, area_flag) %>%
  summarize(count = n())

test <- dataset[dataset$AREA.1920s == "NA",]

# Generate plot
totals<- area_plot3 %>%
    group_by(UD.1920s) %>%
    summarise(total=sum(count))


area_plot <- ggplot(area_plot3, aes(fill=area_flag, y=count, x=UD.1920s)) + 
    geom_bar(position="stack", stat="identity") +
    scale_fill_manual(name = "Area\ndefined", labels = c("no", "yes"), values=c('#999999','darkblue')) +
    geom_text(data=totals, aes(x=UD.1920s, label=total, y=total, fill=NULL), nudge_y=150, size = 3) +
    labs(title = "AREA.1920s", x="Urban district (UD.1920s)", y="Number of data entries")+
    scale_x_continuous(breaks = c(1:21))+
  theme(plot.title = element_text(face = "bold"))
area_plot

3.8 POS.1920s

pos_plot_data <- dataset %>%
  group_by(UD.1920s, POS.1920s) %>%
  summarize(count = n())

pos_plot_data <- data.frame(pos_plot_data, stringsAsFactors = F)

totals<- pos_plot_data  %>%
    group_by(UD.1920s) %>%
    summarise(total=sum(count))

position_pos <- levels(as.factor(pos_plot_data$POS.1920s))

pos_plot <- ggplot(pos_plot_data, aes(fill=factor(POS.1920s, levels = position_pos[c(1,3,5,6,2,4)]), y=count, x=UD.1920s)) + 
    geom_bar(position="stack", stat="identity") +
    scale_fill_manual(name = "Data pattern", labels = c("Data not\navailable", position_pos[c(3,5,6,2,4)]), values = c('#999999','#F0F8FF','#E6E6FA','#87CEFA','#483D8B','darkblue')) +
    geom_text(data=totals, aes(x=UD.1920s, label=total, y=total, fill=NULL), nudge_y=150, size = 3) +
    labs(title = "POS.1920s", x="Urban district (UD.1920s)", y="Number of data entries")+
    scale_x_continuous(breaks = c(1:21))+
  theme(plot.title = element_text(face = "bold"))
pos_plot

3.9 FLOORS.1920s

floors_plot_bez <- dataset %>%
  group_by(UD.1920s, FLOORS.1920s) %>%
  summarize(count = n())

floors_plot_bez <- data.frame(floors_plot_bez, stringsAsFactors = F)


totals<- floors_plot_bez  %>%
    group_by(UD.1920s) %>%
    summarise(total=sum(count))

floor_plot <- ggplot(floors_plot_bez, aes(fill=FLOORS.1920s, y=count, x=UD.1920s)) + 
    geom_bar(position="stack", stat="identity") +
    geom_text(data=totals, aes(x=UD.1920s, label=total, y=total, fill=NULL), nudge_y=150, size = 3) + 
    labs(title = "FLOORS.1920s", x="Urban district (UD.1920s)", y="Number of data entries", fill ="Number of floors\nabove ground floor")  +
    scale_x_continuous(breaks = c(1:21))+
    theme(plot.title = element_text(face = "bold")) +
    geom_point(aes(x = 1, y = 1, size = "Data not\navailable"), shape = NA, colour = "grey") +
    guides(size = guide_legend("", override.aes = list(shape = 15, size = 7)))
floor_plot

## Warning: Using size for a discrete variable is not advised.

## Warning: Removed 114 rows containing missing values (geom_point).

3.10 YoC.1920s

# Create categories

yoc_salz_pos_1 <- grep("^\\d{4}$", dataset$YoC.1920s) # nur 4 stellige Zahlen
yoc_salz_pos_2 <- grep("^\\d{4}[,]", dataset$YoC.1920s) # nur 4 stellige Zahlen am Beginn + ein ,
yoc_salz_pos_3 <- which(dataset$YoC.1920s == "")

dataset$yoc_plot_bez <- rep(NA, nrow(dataset))
dataset[yoc_salz_pos_1, "yoc_plot_bez"] <- "One year date"
dataset[yoc_salz_pos_2, "yoc_plot_bez"] <- "Two year date"
dataset[yoc_salz_pos_3, "yoc_plot_bez"] <- "not available"

# Generate plot

yoc_plot_bez <- dataset %>%
  group_by(UD.1920s, yoc_plot_bez) %>%
  summarize(count = n())

yoc_plot_bez <- data.frame(yoc_plot_bez, stringsAsFactors = F)

totals<- yoc_plot_bez  %>%
    group_by(UD.1920s) %>%
    summarise(total=sum(count))

yoc_plot_bez_fig_factor <- levels(as.factor(yoc_plot_bez$yoc_plot_bez))

yoc_plot <- ggplot(yoc_plot_bez, aes(fill= factor(yoc_plot_bez, levels = yoc_plot_bez_fig_factor[c(1,3,2)]), y=count, x=UD.1920s)) + 
    geom_bar(position="stack", stat="identity") +
    scale_fill_manual(name = "Data pattern", values=c('#999999','lightblue','darkblue')) + 
    geom_text(data=totals, aes(x=UD.1920s, label=total, y=total, fill=NULL), nudge_y=150 , size = 3) +
    labs(title = "YoC.1920s", x="Urban district (UD.1920s)", y="Number of data entries") +
    scale_x_continuous(breaks = c(1:21))+
    theme(plot.title = element_text(face = "bold"))
yoc_plot

3.11 YoP.1920s

# Create categories
yop_salz_pos_1 <- grep("^\\d{4}$", dataset$YoP.1920s) # nur 4 stellige Zahlen
yop_salz_pos_2 <- grep("^\\d{4}[,]", dataset$YoP.1920s) # nur 4 stellige Zahlen am Beginn + ein ,
yop_salz_pos_3 <- which(dataset$YoP.1920s == "")

dataset$yop_plot_bez <- rep(NA, nrow(dataset))
dataset[yop_salz_pos_1, "yop_plot_bez"] <- "One year date"
dataset[yop_salz_pos_2, "yop_plot_bez"] <- "Two year date"
dataset[yop_salz_pos_3, "yop_plot_bez"] <- "not available"

# Generate plot
yop_plot_bez_fig <- dataset %>%
  group_by(UD.1920s, yop_plot_bez) %>%
  summarize(count = n())

yop_plot_bez_fig <- data.frame(yop_plot_bez_fig, stringsAsFactors = F)

totals<- yop_plot_bez_fig  %>%
    group_by(UD.1920s) %>%
    summarise(total=sum(count))

yop_plot_bez_fig_factor <- levels(as.factor(yop_plot_bez_fig$yop_plot_bez))

yop_plot <- ggplot(yop_plot_bez_fig, aes(fill = factor(yop_plot_bez, levels = yop_plot_bez_fig_factor[c(1,3,2)]), y=count, x=UD.1920s)) + 
    geom_bar(position="stack", stat="identity") +
    scale_fill_manual(name = "Data pattern", values=c('#999999','lightblue','darkblue')) +
    geom_text(data=totals, aes(x=UD.1920s, label=total, y=total, fill=NULL), nudge_y=150, size = 3) +
    labs(title="YoP.1920s", x="Urban district (UD.1920s)", y="Number of data entries") +
    scale_x_continuous(breaks = c(1:21))+
  theme(plot.title = element_text(face = "bold"))
yop_plot

## PDF.pages

cadastral$PDF.page.nr <- cadastral$PDF.page.end - cadastral$PDF.page.start + 1

# Page count: Analog building schematic
kg1 <- aggregate(PDF.page.nr ~ UD.1920s, cadastral, sum)

kg12 <- sum(kg1$PDF.page.nr, na.rm = TRUE) # total page number

cadastral$vol_par <- paste(as.character(cadastral$UD.1920s), as.character(cadastral$Volume), as.character(cadastral$Part), sep = "-")

###############

pdf <- unique(dataset[,c("UD.1920s", "Page.pdf")])
pdf <- pdf[order(pdf$UD.1920s, pdf$Page.pdf),]
pdf$Page.pdf <- as.integer(pdf$Page.pdf)


# add volume
volume_raw <- unique(cadastral[,c("Volume", "UD.1920s")])
volume_raw$Volume <- as.integer(volume_raw$Volume)
volume_raw$UD.1920s <- as.integer(volume_raw$UD.1920s)
pdf <- merge(pdf,volume_raw, by.x = "UD.1920s", by.y = "UD.1920s", sort = T)

# Page number per urban district
totals<- kg1[,c("UD.1920s", "PDF.page.nr")]

cadastral$PDF.page.end <- as.integer(cadastral$PDF.page.end)
cadastral$UD.1920s <- as.integer(cadastral$UD.1920s)

max_df <- cadastral %>%
    group_by(UD.1920s) %>%
    summarise(max=max(PDF.page.end))
max_df <- data.frame(max_df, stringsAsFactors = F)
max_df$UD.1920s <- as.integer(max_df$UD.1920s)
max_df$max <- as.integer(max_df$max)
max_df <- max_df[order(max_df$UD.1920s),]

totals$max <- max_df$max
totals$max <- as.numeric(totals$max)
totals$label_text <- paste(rep("[",21),totals$PDF.page.nr, "]",sep = "")

pdf_plot <-ggplot(data=pdf, aes(x=UD.1920s, y=Page.pdf)) +
  geom_point(aes(colour = factor(Volume)), size = 0.1, shape=0) +
  labs(title = "Page.pdf", x="Urban district (UD.1920s)", y="Page number (Page.pdf)\n[page count]") +
  scale_colour_discrete("Volume of analog\nbuilding schematic") +
  guides(color = guide_legend(override.aes = list(size=3, shape = rep(15,10))))+
  theme(plot.title = element_text(face = "bold"))+
  scale_x_continuous(breaks = c(1:21))+
  geom_text(data=totals, aes(x=UD.1920s, label= label_text, y = max), nudge_y=10, size =3)
pdf_plot

4 External validation

4.1 Data Completness: Number of buildings by urban district

This code section produces Figure 7 in the Data Descriptor.

a <- which(dataset$BN.1920s != "") # Einträge mit BN.1920s
b <- which(dataset$BN.1920s == "")

c <- dataset[a,] %>%
  group_by(UD.1920s) %>%
  summarize(n())
c <- data.frame(c)
colnames(c) <- c("UD.1920s", "counts")
c1 <- aggregate(counts ~ UD.1920s, c, sum)

cd2 <- cbind.data.frame(c1, yearbook_1923$Häuser_2) # Es werden die Daten vom Statischtischen Jahrbuch genommen.https://www.digital.wienbibliothek.at/wbrobv/periodical/pageview/2176992
cd2$diff.abs <- cd2[,2]-cd2[,3]
cd2$diff.rel <- round(cd2[,4]/cd2[,3],2)
colnames(cd2) <- c("UD.1920s", "counts.salzberg", "counts.stat", "diff.abs", "diff.rel")

cd2.sum <- data.frame(sum(cd2$counts.salzberg), sum(cd2$counts.stat), sum(cd2$diff.abs), (sum(cd2$diff.abs) / sum(cd2$counts.stat)))

# Plotting figure

data.salz <- data.frame(c(1:21), rep("building schematic", 21), cd2[,2])
data.stat <- data.frame(c(1:21), rep("census", 21), cd2[,3])
coln <- c("UD.1920s", "data_source", "counts")
colnames(data.salz) <- coln
colnames(data.stat) <- coln

data <- rbind(data.stat, data.salz)

p <- ggplot(data, aes(fill=data_source, y=counts, x=UD.1920s)) + 
    geom_bar(position="dodge", stat="identity") +
    ggtitle("Comparative building counts") + 
    scale_x_continuous(labels = c(1:21), breaks = c(1:21)) +
    labs(y= "Number of buildings", x = "Urban district") +
    scale_fill_discrete(name = "Data source", labels = c("Statistical yearbook (1923)", "Digital building schematic (1927-30)")) + 
    theme(legend.position = c(0.23,0.85))
p

4.2 Data plausibility: Number of buildings by floor counts

This code section produces Figure 8 in the Data Descriptor.

floors_salz <- dataset %>%
  group_by(FLOORS.1920s) %>%
  summarise(anz = n())

floors_salz$cum <- cumsum(floors_salz$anz)

# Validierung mit Stockwerksstatitisk von 1914

# Gebäude filtern (Integer, Integer and letter, letter and integer)
buildings_w_bn_pos <- which(dataset$bn.table.ud != "Data not available")

floors <- dataset[buildings_w_bn_pos,] %>%
  group_by(FLOORS.1920s) %>%
  summarize(n_salz = n())

floors$FLOORS.1920s <- as.character(floors$FLOORS.1920s)

floors[7,1] <- "unknown" # Mache NA zu "unknown""

# Gebäude nach Stockwerken 1914 einlesen

floors_1914_wien <- data.frame(colSums(floors_1914[,2:8]))

floors_complete_1914 <- cbind(c("ground floor only", "1", "2", "3", "4", "5 or more", "unknown"), rep("Statistical yearbook (1914)",7), floors_1914_wien$colSums.floors_1914...2.8..)
colnames(floors_complete_1914) <- c("Floors", "data_source", "count")                                                                                          

floors_complete_1920s <- cbind(c("ground floor only", "1", "2", "3", "4", "5 or more", "unknown"), rep("Digital building schematic (1927-30)",7), c(NA, floors[1,2],floors[2,2],floors[3,2],floors[4,2], sum(floors[5:6,2]), floors[7,2]))


colnames(floors_complete_1920s) <- c("Floors","data_source", "count")             
                                                                                                                      
floors_complete <- data.frame(rbind(floors_complete_1914, floors_complete_1920s), stringsAsFactors = F)

rownames(floors_complete) <- NULL
floors_complete$count <- as.numeric(floors_complete$count)
floors_complete$Floors <- as.character(floors_complete$Floors)

positions <- c("ground floor only", "1", "2", "3", "4", "5 or more", "unknown")
floors_complete$count_label <- floors_complete$count
floors_complete$count_label <- as.character(floors_complete$count_label)

floors_complete[8,3] <- 0
floors_complete[8,4] <- "NA"

# plot
floor_valid_plot <- ggplot(floors_complete, aes(x=Floors, y=count, fill=factor(data_source, levels = c("Statistical yearbook (1914)", "Digital building schematic (1927-30)"))))+
  geom_bar(position="dodge", stat="identity")+
  scale_x_discrete(limits = positions)+  
  ggtitle("Comparative building counts") + 
  labs(y= "Number of buildings", x = "Floors") +
  scale_fill_discrete(name = "Data source", labels = c("Statistical yearbook (1914)", "Digital building schematic (1927-30)")) +   
  theme(legend.position = c(0.23,0.85))+
  coord_cartesian(ylim = c(0, 15000))+
  geom_text(aes(label=count_label),  position = position_dodge(0.9), vjust=-1, size = 3)
floor_valid_plot

Technical Validation

Ulrich Kral

14 10 2020