Greek Housing Race: Rents, House Prices and Affordability in Europe

What is this?

The OECD publishes yearly analytical house price indicators for its members and partner countries: an index of rent prices, an index of house prices, and two affordability ratios — price-to-income and price-to-rent. The OECD gives them as indices with 2015 = 100; here every value is turned into a plain percentage change since 2015 — “+13%” means rents are 13% higher than in 2015. They show how much each market moved, not the level of rents in euros.

This vignette pulls the series straight from the OECD data API and builds the same kind of visualisations as the EU Salary Race: four ranked “bar chart races” (one per indicator) and a shareable GIF of each — with Greece highlighted in red against the rest of Europe in gray. It starts with a look at Greece on its own, since rents and house prices there have taken very different paths, followed by the same chart for every other country.

Packages

Code

# install.packages(c("dplyr", "tidyr", "countrycode", "plotly",
#                     "gganimate", "gifski", "scales", "ragg"))

library(dplyr)
library(tidyr)
library(countrycode)  # ISO3 codes -> country names
library(plotly)       # interactive charts
library(gganimate)    # animation
library(gifski)       # GIF renderer
library(scales)       # axis formatting

The data

The OECD serves its datasets through an SDMX API that can answer with a plain CSV, so read.csv() is all we need — no API key, no extra package. The dataset is Analytical house prices indicators (DSD_AN_HOUSE_PRICES@DF_HOUSE_PRICES). We keep the annual observations of the 29 European countries the OECD covers (EU members, Norway, Switzerland, Iceland and the United Kingdom).

Code

url <- paste0("https://sdmx.oecd.org/public/rest/data/",
              "OECD.ECO.MPD,DSD_AN_HOUSE_PRICES@DF_HOUSE_PRICES,1.0/all",
              "?startPeriod=2000&dimensionAtObservation=AllDimensions",
              "&format=csvfilewithlabels")
df_raw <- read.csv(url, fileEncoding = "UTF-8-BOM", check.names = FALSE)

df_raw$STRUCTURE_NAME<-NULL
df_raw$STRUCTURE<-NULL
df_raw$STRUCTURE_ID<-NULL
df_raw$ACTION<-NULL
df_raw$`Time period`<-NULL
df_raw$`Observation value`<-NULL
df_raw$`Base period`<-NULL
df_raw$OBS_STATUS<-NULL
df_raw$Decimals<-NULL
df_raw$DECIMALS<-NULL
df_raw$ADJUSTMENT<-NULL
df_raw$Adjustment<-NULL
df_raw$`Unit multiplier`<-NULL
df_raw$UNIT_MULT<-NULL
df_raw$OBS_STATUS<-NULL
df_raw$`Observation status`<-NULL

europe <- c("AUT", "BEL", "BGR", "HRV", "CZE", "DNK", "EST", "FIN", "FRA", "DEU",
            "GRC", "HUN", "IRL", "ITA", "LVA", "LTU", "LUX", "NLD", "POL", "PRT",
            "ROU", "SVK", "SVN", "ESP", "SWE", "NOR", "CHE", "ISL", "GBR")

df_housing <- df_raw %>%
  filter(FREQ == "A", REF_AREA %in% europe,
         MEASURE %in% c("RPI", "HPI", "HPI_YDH", "HPI_RPI")) %>%
  transmute(iso     = REF_AREA,
            country = countrycode(REF_AREA, origin = "iso3c", destination = "country.name"),
            measure = MEASURE,
            year    = as.integer(TIME_PERIOD),
            values  = round(as.numeric(OBS_VALUE), 1),
            change  = round(values - 100, 1)) %>%  # % change since 2015
  filter(!is.na(values))

head(df_housing)

  iso country measure year values change
1 BEL Belgium HPI_YDH 2000   68.8  -31.2
2 BEL Belgium HPI_YDH 2001   69.0  -31.0
3 BEL Belgium HPI_YDH 2002   72.9  -27.1
4 BEL Belgium HPI_YDH 2003   77.2  -22.8
5 BEL Belgium HPI_YDH 2004   82.5  -17.5
6 BEL Belgium HPI_YDH 2005   89.2  -10.8

Each row is one country, one indicator, one year:

Column	Description
`country`	Country name (derived from the OECD’s ISO3 `REF_AREA` code)
`measure`	`RPI` rent prices, `HPI` nominal house prices, `HPI_YDH` price-to-income ratio, `HPI_RPI` price-to-rent ratio
`year`	Year of observation
`values`	Index, 2015 = 100 (as published)
`change`	Percentage change since 2015 (`values - 100`)

For the races we need one indicator at a time, from 2015 (where every country starts at 0%) onwards, ranked within each year, with Greece in red. A small helper does that for any of the four measures:

Code

race_data <- function(df, which, highlight = "Greece", from = 2015) {
  d <- df[df$measure == which & df$year >= from, ]
  d$rank <- ave(-d$change, d$year, FUN = function(v) rank(v, ties.method = "first"))
  d$color <- ifelse(d$country %in% highlight, "#E74C3C", "gray70")
  d
}

df_rent        <- race_data(df_housing, "RPI")      # rent prices
df_house_price <- race_data(df_housing, "HPI")      # house prices
df_income      <- race_data(df_housing, "HPI_YDH")  # house price to income
df_price_rent  <- race_data(df_housing, "HPI_RPI")  # house price to rent

title_rent        <- "How much more expensive rents got since 2015 (%)"
title_house_price <- "How much more expensive houses got since 2015 (%)"
title_price_rent  <- "How much faster house prices grew than rents since 2015 (%)"
# "+13%" / "-3%"; rounding first (+ 0 turns -0 into 0) so -0.4 reads "+0%", never "-0%"
pct <- function(v) sprintf("%+.0f%%", round(v) + 0)
title_income      <- "How much faster house prices grew than incomes since 2015 (%)"

# data source as a footnote inside each plotly chart (it stays on the chart
# when it is downloaded as a PNG from the plotly toolbar): `b` is the bottom
# margin that makes room for it; the note sits at the bottom of that margin,
# in pixels, so it lands in the same place whatever the chart's height
oecd_url <- paste0("https://data-explorer.oecd.org/vis?df[ds]=dsDisseminateFinalDMZ",
                   "&df[id]=DSD_AN_HOUSE_PRICES%40DF_HOUSE_PRICES&df[ag]=OECD.ECO.MPD")
# chart height in pixels, set on the widget itself (width stays 100% of the
# page), so the chart always fills the box it is drawn in — no scroll bars.
# Returned inside a tagList: printed straight from a chunk, a widget's height
# is overridden by the chunk's figure size; inside a tagList it is kept
# The plotly toolbar goes down the right edge, so it never covers a long title
sized <- function(p, h) {
  p <- plotly::layout(p, modebar = list(orientation = "v"))
  p$height <- h
  htmltools::tagList(p)
}

add_source <- function(p, what, b) {
  plotly::layout(p,
    margin = list(b = b),
    annotations = list(list(
      text = paste0("Source: <a href='", oecd_url, "'>OECD, Analytical house prices indicators</a>",
                    " (", what, ")"),
      xref = "paper", yref = "paper", x = 1, y = 0, yshift = -(b - 22),
      xanchor = "right", yanchor = "top", showarrow = FALSE,
      font = list(size = 11, color = "gray50"))))
}

Greece at a glance

Before racing Greece against Europe, it helps to see its four indicators side by side, as a change against 2015 (the dotted zero line). House prices and rents both fell hard in the years of the crisis, but they recovered at very different speeds: the price-to-rent line shows how much faster house prices have climbed than rents, and — because incomes recovered more slowly still — the price-to-income line shows how much less affordable buying a home has become.

Code

labels_measure <- c(RPI = "Rent prices", HPI = "House prices",
                    HPI_YDH = "Price to income", HPI_RPI = "Price to rent")

df_greece <- df_housing %>%
  filter(country == "Greece", measure %in% names(labels_measure)) %>%
  mutate(indicator = labels_measure[measure])

df_greece_temp <- df_greece[order(df_greece$measure, df_greece$year), c("year", "indicator", "change")]

plot_ly(df_greece_temp,
        x = ~year,
        y = ~change,
        color = ~indicator,
        colors = c("#0082c8", "#E74C3C", "#3cb44b", "#f39c12"),
        type = "scatter",
        mode = "lines+markers",
        text = ~paste0("Year: ", year, "<br>", indicator, ": ", sprintf("%+.1f%%", change)),
        hoverinfo = "text",
        marker = list(size = 10)) %>%
  plotly::layout(title = list(text = "Greece: change in house pricing since 2015 (%)",
                              x = 0.02, xanchor = "left",
                              y = 0.97, yanchor = "top"),
                 margin = list(t = 80),
                 xaxis = list(title = "Year", tickangle = -90),
                 yaxis = list(title = "% vs 2015", ticksuffix = "%"),
                 legend = list(orientation = "h", y = -0.2),
                 font = list(size = 13, color = "gray25"),
                 shapes = list(list(type = "line",
                                    x0 = min(df_greece_temp$year),
                                    x1 = max(df_greece_temp$year),
                                    y0 = 0, y1 = 0,
                                    line = list(dash = "dot", color = "gray")))) %>%
  add_source("RPI, HPI, HPI_YDH, HPI_RPI", b = 150) %>%
  sized(520)

The code below generates a plot for each country, showing the change in rents, house prices, price-to-income and price-to-rent since 2015. Greece is plotted first, then every other country alphabetically.

Code

labels_measure <- c(RPI = "Rent prices", HPI = "House prices",
                    HPI_YDH = "Price to income", HPI_RPI = "Price to rent")

country_plot <- function(cname) {
  d <- df_housing %>%
    filter(country == cname, measure %in% names(labels_measure)) %>%
    mutate(indicator = labels_measure[measure]) %>%
    arrange(measure, year) %>%
    select(year, indicator, change)

  plot_ly(d,
          x = ~year,
          y = ~change,
          color = ~indicator,
          colors = c("#0082c8", "#E74C3C", "#3cb44b", "#f39c12"),
          type = "scatter",
          mode = "lines+markers",
          text = ~paste0("Year: ", year, "<br>", indicator, ": ", sprintf("%+.1f%%", change)),
          hoverinfo = "text",
          marker = list(size = 8)) %>%
    plotly::layout(title = list(text = paste0(cname, ": change in house pricing since 2015 (%)"),
                                x = 0.02, xanchor = "left",
                                y = 0.97, yanchor = "top"),
                   margin = list(t = 80),
                   xaxis = list(title = "", tickangle = -90),
                   yaxis = list(title = "% vs 2015", ticksuffix = "%"),
                   legend = list(orientation = "h", y = -0.2),
                   font = list(size = 13, color = "gray25"),
                   shapes = list(list(type = "line",
                                      x0 = min(d$year), x1 = max(d$year),
                                      y0 = 0, y1 = 0,
                                      line = list(dash = "dot", color = "gray")))) %>%
    add_source("RPI, HPI, HPI_YDH, HPI_RPI", b = 150) %>%
    sized(520)
}

# Greece first, then every other country alphabetically
countries <- c("Greece", sort(setdiff(unique(df_housing$country), "Greece")))

htmltools::tagList(lapply(countries, country_plot))

Building the visualisations

Step 1 — bar chart race, ranked

The frame argument of plot_ly turns a bar chart into a play/pause animation stepping through each year, and plotting by rank instead of country name makes the bars physically reorder every year. Greece is layered on top in red so it is always visible against the gray field. The same function draws four races: rents, house prices, and house prices relative to incomes and to rents.

Code

bar_race <- function(d, title, what) {
  plot_ly() %>%
    add_bars(data = d[d$color == "gray70", ],
             x = ~rank, y = ~change, frame = ~year, ids = ~rank,
             marker = list(color = "gray70"),
             text = ~paste0(country, " ", sprintf("%+.0f%%", change)), hoverinfo = "text",
             textposition = "outside", textangle = 90, showlegend = FALSE) %>%
    add_bars(data = d[d$color %in% "#E74C3C", ],
             x = ~rank, y = ~change, frame = ~year, ids = ~rank,
             marker = list(color = "#E74C3C"),
             text = ~paste0(country, " ", sprintf("%+.0f%%", change)), hoverinfo = "text",
             textposition = "outside", textangle = 90, showlegend = FALSE) %>%
    layout(xaxis  = list(title = "Rank", tickmode = "array"),
           yaxis  = list(title = "% vs 2015", ticksuffix = "%"),
           title  = list(text = title, x = 0.02, xanchor = "left", y = 0.97, yanchor = "top"),
           margin = list(l = 50, r = 30, t = 70),
           font   = list(size = 13, color = "gray25")) %>%
    animation_opts(frame = 1000, transition = 0, easing = "linear", redraw = TRUE) %>%
    add_source(what, b = 200) %>%
    sized(620)
}

bar_race(df_rent, title_rent, "rent prices, RPI")

Code

bar_race(df_house_price, title_house_price, "house prices, HPI")

Code

bar_race(df_income, title_income, "price-to-income ratio, HPI_YDH")

Code

bar_race(df_price_rent, title_price_rent, "price-to-rent ratio, HPI_RPI")

Step 2 — dark themed GIF with `gganimate`

A rendered GIF is easier to share than an interactive chart. ggplot2 + gganimate recreates each of the four ranked races as a horizontal bar chart with a dark background, country labels and a large year watermark.

Code

race_gif <- function(d, title, file) {
  n_countries <- max(d$rank)
  hi  <- max(d$change)
  lo  <- min(0, min(d$change))   # a few markets were below 2015 for a while
  pad <- (hi - lo) * 0.02

  p <- ggplot(d, aes(x = rank, y = change, fill = color, group = country)) +
    geom_col(width = 0.85, show.legend = FALSE) +
    # country names left of the panel, value at the end of each bar
    geom_text(aes(y = lo - pad, label = country, color = color),
              hjust = 1, size = 3.8, fontface = "bold", show.legend = FALSE) +
    geom_text(aes(y = pmax(change, 0) + pad, label = pct(change)),
              hjust = 0, color = "gray80", size = 3.2) +
    geom_hline(yintercept = 0, color = "#546E7A", linewidth = 0.4) +
    scale_fill_identity() +
    scale_color_identity() +
    scale_x_reverse(breaks = NULL) +
    scale_y_continuous(labels = function(v) paste0(v, "%"),
                       expand = expansion(mult = c(0.02, 0.1))) +
    geom_text(
      data = d %>% dplyr::distinct(year),
      aes(x = n_countries, y = hi, label = year),
      inherit.aes = FALSE,
      hjust = 1, vjust = 0,
      size = 20, color = "white", alpha = 0.15,
      fontface = "bold"
    ) +
    coord_flip(clip = "off") +
    labs(title   = title,
         x       = NULL,
         y       = "% vs 2015",
         caption = "Source: OECD Analytical house prices indicators  |  Greece highlighted in red") +
    theme_minimal(base_size = 10) +
    theme(
      plot.background    = element_rect(fill = "#0F1923", color = NA),
      panel.background   = element_rect(fill = "#0F1923", color = NA),
      panel.grid.major.x = element_line(color = "#1E2D3D", linewidth = 0.4),
      panel.grid.major.y = element_blank(),
      panel.grid.minor   = element_blank(),
      axis.text.x   = element_text(color = "#90A4AE", size = 10),
      axis.text.y   = element_blank(),
      axis.title.x  = element_text(color = "#90A4AE", size = 10),
      plot.title    = element_text(color = "white", face = "bold", size = 12,
                                    margin = margin(12, 0, 8, 0)),
      plot.caption  = element_text(color = "#546E7A", size = 8, hjust = 1,
                                    margin = margin(8, 0, 8, 0)),
      plot.margin   = margin(10, 40, 10, 130)) +
    transition_states(year, transition_length = 0, state_length = 5) +
    ease_aes("cubic-in-out")

  animate(p,
          nframes  = 100,
          fps      = 10,
          duration = 20,
          width    = 900,
          height   = 900,
          units    = "px",
          res      = 100,   # set explicitly: when knitting, gganimate otherwise uses the chunk dpi
          renderer = gifski_renderer(file))
}

gifs <- c(rent        = "greek_housing_race_rent.gif",
          house_price = "greek_housing_race_house_price.gif",
          income      = "greek_housing_race_income.gif",
          price_rent  = "greek_housing_race_price_rent.gif")

# each GIF is drawn in its own chunk below: knitr shows only one animate()
# result per chunk

Rents

Code

race_gif(df_rent, title_rent, gifs[["rent"]])

House prices

Code

race_gif(df_house_price, title_house_price, gifs[["house_price"]])

House prices against incomes

Code

race_gif(df_income, title_income, gifs[["income"]])

House prices against rents

Code

race_gif(df_price_rent, title_price_rent, gifs[["price_rent"]])

Note: rendering the four GIFs takes several minutes and requires an internet connection to reach the OECD API. The GIFs are written to the working directory (getwd()).

What the charts show

In Greece, rents kept falling after 2015 and bottomed out in 2018 (-8% against 2015). By 2025 they were +13%, while house prices were +79%: since 2015, buying has become far more expensive than renting.
House prices also outran incomes: the price-to-income ratio was +18% against 2015, which puts Greece 9th of 28 European countries in how much less affordable homes have become.
In the house price race Greece was 16th of 29, and in the price-to-rent race 9th of 27 (+58%): since 2015, house prices in Greece have outgrown rents by that much.
In the rent race Greece was 25th of 27 in 2025: rents rose less there than in most of Europe. The race ranks how much each market moved since 2015, not which country has the most expensive rents.
A caveat for Greece: the OECD rent index comes from the rents in consumer price statistics, which follow all tenancies, most of them on older contracts. Rents asked for new leases, the ones a person moving house actually faces, have risen much faster than this index shows — listing-based indices (e.g. Spitogatos or the Bank of Greece’s analyses) are the place to look for those.

Customising it

Highlight a different country
Codedf_rent <- race_data(df_housing, "RPI", highlight = "Portugal")

Start the race from another year

Code

# e.g. since the pandemic; values stay "% vs 2015" (the OECD's base year)
bar_race(race_data(df_housing, "RPI", from = 2020), title_rent, "rent prices, RPI")

Change the animation speed

The GIFs’ timing is set by the animate() call at the end of race_gif():

Code

animate(p, nframes = 50, fps = 5, duration = 10, ...)   # faster / shorter
animate(p, nframes = 200, fps = 20, duration = 30, ...) # smoother / longer

OECD Analytical house prices indicators — the dataset in the OECD Data Explorer
OECD Affordable Housing Database — housing costs as a share of income, tenure and social housing indicators
OECD SDMX API — how to build data queries
gganimate documentation — all transition types and rendering options

Rendered with R 4.6.1 · plotly · ggplot2 · gganimate · OECD SDMX API

--- title: "Greek Housing Race: Rents, House Prices and Affordability in Europe" author: "" output: html_document: css: styles.css theme: flatly highlight: tango toc: true toc_float: true code_folding: show self_contained: true vignette: > %\VignetteIndexEntry{Greek Housing Race} %\VignetteEngine{knitr::rmarkdown} %\VignetteEncoding{UTF-8} --- ```{r setup, include=FALSE} directory<-if (rstudioapi::isAvailable()) paste0(dirname(rstudioapi::getActiveDocumentContext()$path),"/") else "" source(paste0(directory, "R/code_options.R")) source(paste0(directory, "R/code_rwf.R")) ``` ## What is this? The OECD publishes yearly **analytical house price indicators** for its members and partner countries: an index of **rent prices**, an index of **house prices**, and two affordability ratios — **price-to-income** and **price-to-rent**. The OECD gives them as indices with 2015 = 100; here every value is turned into a plain **percentage change since 2015** — "+13%" means rents are 13% higher than in 2015. They show how much each market moved, not the level of rents in euros. This vignette pulls the series straight from the OECD data API and builds the same kind of visualisations as the [EU Salary Race](ILLUSTRATION_EU_SALARY.html): four ranked "bar chart races" (one per indicator) and a shareable GIF of each — with Greece highlighted in red against the rest of Europe in gray. It starts with a look at Greece on its own, since rents and house prices there have taken very different paths, followed by the same chart for every other country. --- ## Packages ```{r packages} # install.packages(c("dplyr", "tidyr", "countrycode", "plotly", # "gganimate", "gifski", "scales", "ragg")) library(dplyr) library(tidyr) library(countrycode) # ISO3 codes -> country names library(plotly) # interactive charts library(gganimate) # animation library(gifski) # GIF renderer library(scales) # axis formatting ``` --- ## The data The OECD serves its datasets through an SDMX API that can answer with a plain CSV, so `read.csv()` is all we need — no API key, no extra package. The dataset is *Analytical house prices indicators* (`DSD_AN_HOUSE_PRICES@DF_HOUSE_PRICES`). We keep the **annual** observations of the 29 European countries the OECD covers (EU members, Norway, Switzerland, Iceland and the United Kingdom). ```{r data} url <- paste0("https://sdmx.oecd.org/public/rest/data/", "OECD.ECO.MPD,DSD_AN_HOUSE_PRICES@DF_HOUSE_PRICES,1.0/all", "?startPeriod=2000&dimensionAtObservation=AllDimensions", "&format=csvfilewithlabels") df_raw <- read.csv(url, fileEncoding = "UTF-8-BOM", check.names = FALSE) df_raw$STRUCTURE_NAME<-NULL df_raw$STRUCTURE<-NULL df_raw$STRUCTURE_ID<-NULL df_raw$ACTION<-NULL df_raw$`Time period`<-NULL df_raw$`Observation value`<-NULL df_raw$`Base period`<-NULL df_raw$OBS_STATUS<-NULL df_raw$Decimals<-NULL df_raw$DECIMALS<-NULL df_raw$ADJUSTMENT<-NULL df_raw$Adjustment<-NULL df_raw$`Unit multiplier`<-NULL df_raw$UNIT_MULT<-NULL df_raw$OBS_STATUS<-NULL df_raw$`Observation status`<-NULL europe <- c("AUT", "BEL", "BGR", "HRV", "CZE", "DNK", "EST", "FIN", "FRA", "DEU", "GRC", "HUN", "IRL", "ITA", "LVA", "LTU", "LUX", "NLD", "POL", "PRT", "ROU", "SVK", "SVN", "ESP", "SWE", "NOR", "CHE", "ISL", "GBR") df_housing <- df_raw %>% filter(FREQ == "A", REF_AREA %in% europe, MEASURE %in% c("RPI", "HPI", "HPI_YDH", "HPI_RPI")) %>% transmute(iso = REF_AREA, country = countrycode(REF_AREA, origin = "iso3c", destination = "country.name"), measure = MEASURE, year = as.integer(TIME_PERIOD), values = round(as.numeric(OBS_VALUE), 1), change = round(values - 100, 1)) %>% # % change since 2015 filter(!is.na(values)) head(df_housing) ``` Each row is one country, one indicator, one year: | Column | Description | |--------|-------------| | `country` | Country name (derived from the OECD's ISO3 `REF_AREA` code) | | `measure` | `RPI` rent prices, `HPI` nominal house prices, `HPI_YDH` price-to-income ratio, `HPI_RPI` price-to-rent ratio | | `year` | Year of observation | | `values` | Index, 2015 = 100 (as published) | | `change` | Percentage change since 2015 (`values - 100`) | For the races we need one indicator at a time, from 2015 (where every country starts at 0%) onwards, ranked within each year, with Greece in red. A small helper does that for any of the four measures: ```{r race_data} race_data <- function(df, which, highlight = "Greece", from = 2015) { d <- df[df$measure == which & df$year >= from, ] d$rank <- ave(-d$change, d$year, FUN = function(v) rank(v, ties.method = "first")) d$color <- ifelse(d$country %in% highlight, "#E74C3C", "gray70") d } df_rent <- race_data(df_housing, "RPI") # rent prices df_house_price <- race_data(df_housing, "HPI") # house prices df_income <- race_data(df_housing, "HPI_YDH") # house price to income df_price_rent <- race_data(df_housing, "HPI_RPI") # house price to rent title_rent <- "How much more expensive rents got since 2015 (%)" title_house_price <- "How much more expensive houses got since 2015 (%)" title_price_rent <- "How much faster house prices grew than rents since 2015 (%)" # "+13%" / "-3%"; rounding first (+ 0 turns -0 into 0) so -0.4 reads "+0%", never "-0%" pct <- function(v) sprintf("%+.0f%%", round(v) + 0) title_income <- "How much faster house prices grew than incomes since 2015 (%)" # data source as a footnote inside each plotly chart (it stays on the chart # when it is downloaded as a PNG from the plotly toolbar): `b` is the bottom # margin that makes room for it; the note sits at the bottom of that margin, # in pixels, so it lands in the same place whatever the chart's height oecd_url <- paste0("https://data-explorer.oecd.org/vis?df[ds]=dsDisseminateFinalDMZ", "&df[id]=DSD_AN_HOUSE_PRICES%40DF_HOUSE_PRICES&df[ag]=OECD.ECO.MPD") # chart height in pixels, set on the widget itself (width stays 100% of the # page), so the chart always fills the box it is drawn in — no scroll bars. # Returned inside a tagList: printed straight from a chunk, a widget's height # is overridden by the chunk's figure size; inside a tagList it is kept # The plotly toolbar goes down the right edge, so it never covers a long title sized <- function(p, h) { p <- plotly::layout(p, modebar = list(orientation = "v")) p$height <- h htmltools::tagList(p) } add_source <- function(p, what, b) { plotly::layout(p, margin = list(b = b), annotations = list(list( text = paste0("Source: <a href='", oecd_url, "'>OECD, Analytical house prices indicators</a>", " (", what, ")"), xref = "paper", yref = "paper", x = 1, y = 0, yshift = -(b - 22), xanchor = "right", yanchor = "top", showarrow = FALSE, font = list(size = 11, color = "gray50")))) } ``` --- ## Greece at a glance Before racing Greece against Europe, it helps to see its four indicators side by side, as a change against 2015 (the dotted zero line). House prices and rents both fell hard in the years of the crisis, but they recovered at very different speeds: the price-to-rent line shows how much faster house prices have climbed than rents, and — because incomes recovered more slowly still — the price-to-income line shows how much less affordable buying a home has become. ```{r greece} labels_measure <- c(RPI = "Rent prices", HPI = "House prices", HPI_YDH = "Price to income", HPI_RPI = "Price to rent") df_greece <- df_housing %>% filter(country == "Greece", measure %in% names(labels_measure)) %>% mutate(indicator = labels_measure[measure]) df_greece_temp <- df_greece[order(df_greece$measure, df_greece$year), c("year", "indicator", "change")] plot_ly(df_greece_temp, x = ~year, y = ~change, color = ~indicator, colors = c("#0082c8", "#E74C3C", "#3cb44b", "#f39c12"), type = "scatter", mode = "lines+markers", text = ~paste0("Year: ", year, "<br>", indicator, ": ", sprintf("%+.1f%%", change)), hoverinfo = "text", marker = list(size = 10)) %>% plotly::layout(title = list(text = "Greece: change in house pricing since 2015 (%)", x = 0.02, xanchor = "left", y = 0.97, yanchor = "top"), margin = list(t = 80), xaxis = list(title = "Year", tickangle = -90), yaxis = list(title = "% vs 2015", ticksuffix = "%"), legend = list(orientation = "h", y = -0.2), font = list(size = 13, color = "gray25"), shapes = list(list(type = "line", x0 = min(df_greece_temp$year), x1 = max(df_greece_temp$year), y0 = 0, y1 = 0, line = list(dash = "dot", color = "gray")))) %>% add_source("RPI, HPI, HPI_YDH, HPI_RPI", b = 150) %>% sized(520) ``` The code below generates a plot for each country, showing the change in rents, house prices, price-to-income and price-to-rent since 2015. Greece is plotted first, then every other country alphabetically. ```{r all_countries} labels_measure <- c(RPI = "Rent prices", HPI = "House prices", HPI_YDH = "Price to income", HPI_RPI = "Price to rent") country_plot <- function(cname) { d <- df_housing %>% filter(country == cname, measure %in% names(labels_measure)) %>% mutate(indicator = labels_measure[measure]) %>% arrange(measure, year) %>% select(year, indicator, change) plot_ly(d, x = ~year, y = ~change, color = ~indicator, colors = c("#0082c8", "#E74C3C", "#3cb44b", "#f39c12"), type = "scatter", mode = "lines+markers", text = ~paste0("Year: ", year, "<br>", indicator, ": ", sprintf("%+.1f%%", change)), hoverinfo = "text", marker = list(size = 8)) %>% plotly::layout(title = list(text = paste0(cname, ": change in house pricing since 2015 (%)"), x = 0.02, xanchor = "left", y = 0.97, yanchor = "top"), margin = list(t = 80), xaxis = list(title = "", tickangle = -90), yaxis = list(title = "% vs 2015", ticksuffix = "%"), legend = list(orientation = "h", y = -0.2), font = list(size = 13, color = "gray25"), shapes = list(list(type = "line", x0 = min(d$year), x1 = max(d$year), y0 = 0, y1 = 0, line = list(dash = "dot", color = "gray")))) %>% add_source("RPI, HPI, HPI_YDH, HPI_RPI", b = 150) %>% sized(520) } # Greece first, then every other country alphabetically countries <- c("Greece", sort(setdiff(unique(df_housing$country), "Greece"))) htmltools::tagList(lapply(countries, country_plot)) ``` --- ## Building the visualisations ### Step 1 — bar chart race, ranked The `frame` argument of `plot_ly` turns a bar chart into a play/pause animation stepping through each year, and plotting by `rank` instead of country name makes the bars physically reorder every year. Greece is layered on top in red so it is always visible against the gray field. The same function draws four races: rents, house prices, and house prices relative to incomes and to rents. ```{r bar_race} bar_race <- function(d, title, what) { plot_ly() %>% add_bars(data = d[d$color == "gray70", ], x = ~rank, y = ~change, frame = ~year, ids = ~rank, marker = list(color = "gray70"), text = ~paste0(country, " ", sprintf("%+.0f%%", change)), hoverinfo = "text", textposition = "outside", textangle = 90, showlegend = FALSE) %>% add_bars(data = d[d$color %in% "#E74C3C", ], x = ~rank, y = ~change, frame = ~year, ids = ~rank, marker = list(color = "#E74C3C"), text = ~paste0(country, " ", sprintf("%+.0f%%", change)), hoverinfo = "text", textposition = "outside", textangle = 90, showlegend = FALSE) %>% layout(xaxis = list(title = "Rank", tickmode = "array"), yaxis = list(title = "% vs 2015", ticksuffix = "%"), title = list(text = title, x = 0.02, xanchor = "left", y = 0.97, yanchor = "top"), margin = list(l = 50, r = 30, t = 70), font = list(size = 13, color = "gray25")) %>% animation_opts(frame = 1000, transition = 0, easing = "linear", redraw = TRUE) %>% add_source(what, b = 200) %>% sized(620) } bar_race(df_rent, title_rent, "rent prices, RPI") bar_race(df_house_price, title_house_price, "house prices, HPI") bar_race(df_income, title_income, "price-to-income ratio, HPI_YDH") bar_race(df_price_rent, title_price_rent, "price-to-rent ratio, HPI_RPI") ``` ### Step 2 — dark themed GIF with `gganimate` A rendered GIF is easier to share than an interactive chart. `ggplot2` + `gganimate` recreates each of the four ranked races as a horizontal bar chart with a dark background, country labels and a large year watermark. ```{r gif, eval=TRUE} race_gif <- function(d, title, file) { n_countries <- max(d$rank) hi <- max(d$change) lo <- min(0, min(d$change)) # a few markets were below 2015 for a while pad <- (hi - lo) * 0.02 p <- ggplot(d, aes(x = rank, y = change, fill = color, group = country)) + geom_col(width = 0.85, show.legend = FALSE) + # country names left of the panel, value at the end of each bar geom_text(aes(y = lo - pad, label = country, color = color), hjust = 1, size = 3.8, fontface = "bold", show.legend = FALSE) + geom_text(aes(y = pmax(change, 0) + pad, label = pct(change)), hjust = 0, color = "gray80", size = 3.2) + geom_hline(yintercept = 0, color = "#546E7A", linewidth = 0.4) + scale_fill_identity() + scale_color_identity() + scale_x_reverse(breaks = NULL) + scale_y_continuous(labels = function(v) paste0(v, "%"), expand = expansion(mult = c(0.02, 0.1))) + geom_text( data = d %>% dplyr::distinct(year), aes(x = n_countries, y = hi, label = year), inherit.aes = FALSE, hjust = 1, vjust = 0, size = 20, color = "white", alpha = 0.15, fontface = "bold" ) + coord_flip(clip = "off") + labs(title = title, x = NULL, y = "% vs 2015", caption = "Source: OECD Analytical house prices indicators | Greece highlighted in red") + theme_minimal(base_size = 10) + theme( plot.background = element_rect(fill = "#0F1923", color = NA), panel.background = element_rect(fill = "#0F1923", color = NA), panel.grid.major.x = element_line(color = "#1E2D3D", linewidth = 0.4), panel.grid.major.y = element_blank(), panel.grid.minor = element_blank(), axis.text.x = element_text(color = "#90A4AE", size = 10), axis.text.y = element_blank(), axis.title.x = element_text(color = "#90A4AE", size = 10), plot.title = element_text(color = "white", face = "bold", size = 12, margin = margin(12, 0, 8, 0)), plot.caption = element_text(color = "#546E7A", size = 8, hjust = 1, margin = margin(8, 0, 8, 0)), plot.margin = margin(10, 40, 10, 130)) + transition_states(year, transition_length = 0, state_length = 5) + ease_aes("cubic-in-out") animate(p, nframes = 100, fps = 10, duration = 20, width = 900, height = 900, units = "px", res = 100, # set explicitly: when knitting, gganimate otherwise uses the chunk dpi renderer = gifski_renderer(file)) } gifs <- c(rent = "greek_housing_race_rent.gif", house_price = "greek_housing_race_house_price.gif", income = "greek_housing_race_income.gif", price_rent = "greek_housing_race_price_rent.gif") # each GIF is drawn in its own chunk below: knitr shows only one animate() # result per chunk ``` #### Rents ```{r gif0, eval=TRUE} race_gif(df_rent, title_rent, gifs[["rent"]]) ``` #### House prices ```{r gif1, eval=TRUE} race_gif(df_house_price, title_house_price, gifs[["house_price"]]) ``` #### House prices against incomes ```{r gif2, eval=TRUE} race_gif(df_income, title_income, gifs[["income"]]) ``` #### House prices against rents ```{r gif3, eval=TRUE} race_gif(df_price_rent, title_price_rent, gifs[["price_rent"]]) ``` > **Note:** rendering the four GIFs takes several minutes and requires an internet connection to reach the OECD API. The GIFs are written to the working directory (`getwd()`). --- ## What the charts show ```{r facts, include=FALSE} last_year <- max(df_housing$year) grc <- function(m, y) df_housing$change[df_housing$country == "Greece" & df_housing$measure == m & df_housing$year == y] grc_rent <- df_rent[df_rent$country == "Greece", ] grc_rent_low_year <- grc_rent$year[which.min(grc_rent$change)] grc_rent_rank <- df_rent$rank[df_rent$country == "Greece" & df_rent$year == last_year] grc_income_rank <- df_income$rank[df_income$country == "Greece" & df_income$year == last_year] grc_house_rank <- df_house_price$rank[df_house_price$country == "Greece" & df_house_price$year == last_year] grc_prent_rank <- df_price_rent$rank[df_price_rent$country == "Greece" & df_price_rent$year == last_year] # countries with data in the last year (a few have not published it yet) n_last <- function(d) sum(d$year == last_year) ``` - In Greece, rents kept falling after 2015 and bottomed out in **`r grc_rent_low_year`** (**`r pct(min(grc_rent$change))`** against 2015). By `r last_year` they were **`r pct(grc("RPI", last_year))`**, while house prices were **`r pct(grc("HPI", last_year))`**: since 2015, buying has become far more expensive than renting. - House prices also outran incomes: the price-to-income ratio was **`r pct(grc("HPI_YDH", last_year))`** against 2015, which puts Greece **`r scales::ordinal(grc_income_rank)`** of `r n_last(df_income)` European countries in how much less affordable homes have become. - In the house price race Greece was **`r scales::ordinal(grc_house_rank)`** of `r n_last(df_house_price)`, and in the price-to-rent race **`r scales::ordinal(grc_prent_rank)`** of `r n_last(df_price_rent)` (**`r pct(grc("HPI_RPI", last_year))`**): since 2015, house prices in Greece have outgrown rents by that much. - In the rent race Greece was **`r scales::ordinal(grc_rent_rank)`** of `r n_last(df_rent)` in `r last_year`: rents rose less there than in most of Europe. The race ranks how much each market moved since 2015, not which country has the most expensive rents. - **A caveat for Greece:** the OECD rent index comes from the rents in consumer price statistics, which follow *all* tenancies, most of them on older contracts. Rents asked for new leases, the ones a person moving house actually faces, have risen much faster than this index shows — listing-based indices (e.g. Spitogatos or the Bank of Greece's analyses) are the place to look for those. --- ## Customising it ### Highlight a different country ```{r custom_highlight, eval=FALSE} df_rent <- race_data(df_housing, "RPI", highlight = "Portugal") ``` ### Start the race from another year ```{r custom_from, eval=FALSE} # e.g. since the pandemic; values stay "% vs 2015" (the OECD's base year) bar_race(race_data(df_housing, "RPI", from = 2020), title_rent, "rent prices, RPI") ``` ### Change the animation speed The GIFs' timing is set by the `animate()` call at the end of `race_gif()`: ```{r custom_speed, eval=FALSE} animate(p, nframes = 50, fps = 5, duration = 10, ...) # faster / shorter animate(p, nframes = 200, fps = 20, duration = 30, ...) # smoother / longer ``` --- ## Further reading - [OECD Analytical house prices indicators](https://data-explorer.oecd.org/vis?df[ds]=dsDisseminateFinalDMZ&df[id]=DSD_AN_HOUSE_PRICES%40DF_HOUSE_PRICES&df[ag]=OECD.ECO.MPD) — the dataset in the OECD Data Explorer - [OECD Affordable Housing Database](https://www.oecd.org/en/data/datasets/oecd-affordable-housing-database.html) — housing costs as a share of income, tenure and social housing indicators - [OECD SDMX API](https://gitlab.algobank.oecd.org/public-documentation/dotstat-migration/-/raw/main/OECD_Data_API_documentation.pdf) — how to build data queries - [gganimate documentation](https://gganimate.com/) — all transition types and rendering options --- *Rendered with R `r getRversion()` · plotly · ggplot2 · gganimate · OECD SDMX API*

What is this?

Packages

The data

Greece at a glance

Building the visualisations

Step 1 — bar chart race, ranked

Step 2 — dark themed GIF with gganimate

Rents

House prices

House prices against incomes

House prices against rents

What the charts show

Customising it

Highlight a different country

Start the race from another year

Change the animation speed

Further reading

Step 2 — dark themed GIF with `gganimate`