Some Graphs about COVID-19 in Italy

Table of Contents

Menu

Introduction

This page contains various plots generated from that data using Org Mode and R: no fancy web services, just plain-old off-line generation. On top of being an interesting exercise on R and literate programming in Emacs, I use this page to get an idea of the evolution of the pandemic in Italy.

This page was created on <2020-03-28 Sat> and last updated on <2021-07-17 Sat>.

The source code available on the COVID-19 pages is distributed under the MIT License; the content is distributed under a Creative Commons - Attribution 4.0.

Get the data

First we get the data:

curl https://raw.githubusercontent.com/pcm-dpc/COVID-19/master/dati-regioni/dpc-covid19-ita-regioni.csv > data/dpc-covid19-ita-regioni.csv
curl https://raw.githubusercontent.com/pcm-dpc/COVID-19/master/dati-andamento-nazionale/dpc-covid19-ita-andamento-nazionale.csv > data/dpc-covid19-ita-andamento-nazionale.csv 
curl https://raw.githubusercontent.com/pcm-dpc/COVID-19/master/dati-regioni/dpc-covid19-ita-regioni-latest.csv > data/dpc-covid19-ita-regioni-latest.csv

R Functions

This section contains the code for plotting data. The function my_plot plots different variables of an input dataframe over time, optionally filtering over region, which is the denomination of an Italian region.

The optional argument max defines the maximum value for the x-axis, while the optional Boolean arguments textlabels and filter control, respectively, whether text labels are printed on graphs and data has to be filtered by Region.

Finally, the optional arguments variables, graphtypes, and colors are vectors, defining, respectively, the variables to plot, the type of plot, and the colors used.

my_plot <- function(region, data, max=-1, textlabels=TRUE, filter=FALSE, 
                    variables  = c("totale_casi", "nuovi_positivi", "totale_positivi", "deceduti", "dimessi_guariti"),
                    graphtypes = c("l", "h", "l", "l", "l"),
                    colors     = c("red", "black", "orange", "slategrey", "forestgreen")) {
  par(cex=1.40, las=2)

  # if asked to filter, filter data according to region
  if (filter) {
    dataframe <- subset(data, denominazione_regione == region)
  }
  else {
    dataframe <- data
  }

  if (max == -1) {
    max=max(dataframe$totale_casi)
  }

  plot(x=1, 
       xlim=c(min(data$data), max(data$data)),
       ylim=c(0,max),
       type="n",
       main = region,
       xlab="",
       ylab="",
       xaxt="n")

  axis.Date(1, at=dataframe$data, by="days", format="%b %d")

  # do the plots, now
  for (i in 1:length(variables)) {
    lines(x=dataframe$data, y=dataframe[, variables[i]],
          type=graphtypes[i], 
          lwd=5,
          pch=16, 
          col=colors[i])
    if (textlabels) {
      text(x=dataframe$data, y=dataframe[, variables[i]], 
           label=dataframe[, variables[i]], 
           pos=2, 
           col=colors[i])
    }
  }

  values = sprintf("(%s)", dataframe[nrow(dataframe), variables])
  legend("topleft", legend=paste(variables, values), col=colors, lty=1, cex=1.6)
  grid(col = "lightgray")
}

Then we read the data from the CSV files of the Civil Protection repository:

PATH="./data/"

# evolution over time, by Region
data = read.csv(file.path(PATH, "dpc-covid19-ita-regioni.csv"))
data$data <- as.Date(data$data)

# evolution over time at the National level
national = read.csv(file.path(PATH, "dpc-covid19-ita-andamento-nazionale.csv"))
national$data <- as.Date(national$data)

# latest regional data
latest = read.csv(file.path(PATH, "dpc-covid19-ita-regioni-latest.csv"))
latest$data <- as.Date(national$data)

We are now ready to print and plot the data.

This Week in Italy

cols = c(
  "ricoverati_con_sintomi", 
  "terapia_intensiva",
  "totale_ospedalizzati",
  "isolamento_domiciliare", 
  "totale_positivi",
  "nuovi_positivi",
  "dimessi_guariti",
  "deceduti",
  "totale_casi"
)
labels = c(
  "In hospitals with symptoms", 
  "In ICUs",
  "Total hospitalized",
  "Quarantined at home", 
  "Active cases",
  "New cases",
  "Recovered",
  "Deaths",
  "Total number of cases"
)

Today = unlist(national[nrow(national), cols])
Yesterday = unlist(national[nrow(national) - 1, cols])
TwoDaysAgo = unlist(national[nrow(national) - 2, cols])
ThreeDaysAgo = unlist(national[nrow(national) - 3, cols])
FourDaysAgo = unlist(national[nrow(national) - 4, cols])
FiveDaysAgo = unlist(national[nrow(national) - 5, cols])

output_frame <- data.frame(labels, FiveDaysAgo, FourDaysAgo, ThreeDaysAgo, TwoDaysAgo, Yesterday, Today)
colnames(output_frame) <- rev(seq(Sys.Date(), by="-1 day", length.out=7))
colnames(output_frame)[1] <- "Label"
output_frame
Label 2021-10-05 2021-10-06 2021-10-07 2021-10-08 2021-10-09 2021-10-10
In hospitals with symptoms 3032 2968 2872 2824 2742 2692
In ICUs 437 433 415 403 383 367
Total hospitalized 3469 3401 3287 3227 3125 3059
Quarantined at home 88627 86898 84960 83946 82801 82243
Active cases 92096 90299 88247 87173 85926 85302
New cases 1612 2466 3235 2938 3023 2748
Recovered 4460482 4464692 4469937 4473903 4478137 4481462
Deaths 131068 131118 131157 131198 131228 131274
Total number of cases 4683646 4686109 4689341 4692274 4695291 4698038

Variations with respect to previous day

We now plot the variations in the last week, that is the difference between a day and the previous day. In many cases, the lower the number, the better. In other cases (e.g., Recovered), the higher, the better.

Diff4 = FourDaysAgo - FiveDaysAgo
Diff3 = ThreeDaysAgo - FourDaysAgo
Diff2 = TwoDaysAgo - ThreeDaysAgo 
Diff1 = Yesterday - TwoDaysAgo
Diff0 = Today - Yesterday

diff_frame <- data.frame(labels, Diff4, Diff3, Diff2, Diff1, Diff0)
diff_frame
labels Diff4 Diff3 Diff2 Diff1 Diff0
In hospitals with symptoms -64 -96 -48 -82 -50
In ICUs -4 -18 -12 -20 -16
Total hospitalized -68 -114 -60 -102 -66
Quarantined at home -1729 -1938 -1014 -1145 -558
Active cases -1797 -2052 -1074 -1247 -624
New cases 854 769 -297 85 -275
Recovered 4210 5245 3966 4234 3325
Deaths 50 39 41 30 46
Total number of cases 2463 3232 2933 3017 2747

See also the historical series of new cases in Italy.

Situation in Italy

Overall Situation

Evolution over time.

my_plot("Italia", national, textlabels=FALSE)

italia.png

Breakdown of Quarantine

It tells where people with COVID-19 are spending their quarantine, that is, a breakdown of the “yellow” line of the previous plot.

The blue line is the number of people hospedalized during the (first) lockdown. Now the capacity of the health system should be higher, but it seems something to look at (although the situation differs from region to region).

  my_plot("Italia", 
          national,
          max(national$isolamento_domiciliare), textlabels=FALSE, filter=FALSE, 
          variables=c("ricoverati_con_sintomi", "terapia_intensiva", "totale_ospedalizzati", "isolamento_domiciliare"),  
          graphtypes=c("l", "l", "l", "l", "h"),
          colors=c("#FECEAB", "#EC2049", "#E84A5F", "#A7226E"))
abline(h = national[38,]$totale_ospedalizzati, col="#330000", lwd=2, lty=3)
abline(v = as.Date("2020-10-25"), col="#330000", lwd=2, lty=3)

hospitalized.png

Focus on Trentino, Liguria, Veneto and Lombardia

Situation in Trentino

my_plot("P.A. Trento", data, filter=TRUE, textlabels=FALSE)

trentino.png

Situation in Liguria

my_plot("Liguria", data, filter=TRUE, textlabels=FALSE)

liguria.png

Situation in Veneto

my_plot("Veneto", data, filter=TRUE, textlabels=FALSE)

veneto.png

Situation in Lombardia

my_plot("Lombardia", data, filter=TRUE, textlabels=FALSE)

lombardia.png

Situation by Region

Situation by Region

# how many rows and columns?
par(mfrow=c(11, 2))

max <- max(data$totale_casi)

regions <- c("Valle d'Aosta", "Piemonte", "Liguria", "Lombardia", "Veneto",
             "P.A. Trento", "P.A. Bolzano", "Friuli Venezia Giulia",
             "Emilia-Romagna", "Toscana", "Marche", "Umbria",
             "Lazio", "Abruzzo", "Molise", "Campania",
             "Puglia", "Basilicata", "Calabria", "Sicilia",
             "Sardegna")
for (region in regions) {
  my_plot(
    region, data, filter=TRUE, textlabels=FALSE,
          variables=c("totale_casi", "totale_positivi", "deceduti", "dimessi_guariti"),
          max = max,
          graphtypes=c("l", "l", "l", "l"),
          colors=c("red", "orange", "slategrey", "forestgreen"))
}

cases_by_region.png