Talk:COVID-19 pandemic in Italy/Archive 4

Archive 1

Archive 2

Archive 3

Archive 4

Archive 5

Graph charts and dates

I recently pushed a revision of the graphical layout of the charts in this section of the article. However this was reverted by Nick.mon because they preferred the older version. I would say that my version has many positive things and only few negative ones. It uses the Template:Graph:Chart feature of the date type for the x axis (which is very appropriate for our charts here), also meaning that we have the great advantage that the template automatically fixes the number of ticks that appear on the axis. At the moment the main problem with the current graph is that we cannot squeeze the data so much otherwise the date names on the x axis get on top of each other. This is a huge problem, because in one month we will have a chart that goes beyond the width of a regular browser page. On the other hand, using the date type, the issue of the ticks going onto each other disappears, because only a few ones are printed (e.g. instead of 1 Mar, 2 Mar, 3 Mar, 4 Mar, 5 Mar, it just shows 1 Mar and 5 Mar), so we can keep the width of the graph contained even next month, and this summer. I also switched the plot graphics to the line-type without showing the point at each value, because this way it looks more continuous, one can follow the trend more easily, and especially because in a few weeks the density of points would have become much higher and they would have gone one onto each other, creating a graphical mess.

The main issue with my edit and with the date type was that at the moment it doesn't work well with the rect i.e. the bar graph type, that we are currently using for the "per day" charts. In particular, as I already explained in Template talk:Graph:Chart, there is a bug with the date format that is shown on the x axis (and I hope this bug would be solved soon, if we put a bit of pressure on the editors who are expert there). So that's why I switched to line type for the "per day" chart, and I don't think this is a huge loss, except that the daily values cannot be shown (but they are already there in the tables in the section above). This is a minor complication, I think, but the advantage we gain from the fact that we don't need to increase the width of the graph every 5 days is priceless.

So, while I am open to criticism and corrections, I think we definitely should switch to date type, even though this means (hopefully momentarily, until the bug is fixed) we cannot have the bar chart for the daily cases. Another option would be to keep the daily graphs as they are now and switch the type only in the "cumulative" graphs. --Ritchie92 (talk) 08:07, 17 April 2020 (UTC)

Ritchie92, I must be sincere, I'm not an expert of these templates and your reasons are solid, I only prefer the previous version from an "aesthetic" point of view, and for consistency with the other articles. So, if we don't find another solution, I'm open to use your version. Anyway, shouldn't we use the current one, with the tool "< div ></ div >" so that a reader could slide the graph? -- Nick.mon (talk) 08:25, 17 April 2020 (UTC)

This is what I mean:

Sure, the div environment would be mandatory in the case the graph is too wide, however I think it's a bit exaggerated to have a graph that is longer than an average browser page, for a time span that is of 2-3 months, or a bit more. Actually the pandemic could also last a year or more according to some predictions, so we might find ourselves in a situation were these graphs are as wide as 3-4 full pages widths! This would be very irritating. Also, having a wider graph means that the trends are less visible to a human eye (every trend looks "slowed down"). So I would still try to avoid this situation and have more compact graphs that are visible without having to scroll. --Ritchie92 (talk) 08:38, 17 April 2020 (UTC)

Just as reference, this is what my proposal would look like (equivalent to the graph above) – I show also a squeezed version to show how it would look in the next months

Ok ok, I understand your point of view. Regarding your graphs, in my view, the first one (the one with dots) would be better, but this is only my opinion. Anyway, let's see what other users suggest. -- Nick.mon (talk) 09:14, 17 April 2020 (UTC)

Can anyone add daily test counts to these graphs? --Dan Polansky (talk) 10:10, 19 April 2020 (UTC)

A separate graph with test counts is already in the article. I don't think we should have the test counts on the same graph as the confirmed cases: this would just ruin the visibility of the data because the number of tests performed is much higher than all other numbers. --Ritchie92 (talk) 09:23, 20 April 2020 (UTC)

You would need to rescale the test count and then put it to the same graph; the scale for the test count could be put to the right vertical axis. --Dan Polansky (talk) 16:26, 20 April 2020 (UTC)

And what would be very good to have and what some countries provide on their websites is the daily confirmed case count expressed as a percentage of the daily test count. --Dan Polansky (talk) 16:27, 20 April 2020 (UTC)

the scale for the test count could be put to the right vertical axis this cannot be done with the Template:Graph:Chart that we are using to keep the graph source accessible to editors.

the daily confirmed case count expressed as a percentage of the daily test count this appears in the other template count at the beginning of the page, the vertical one. --Ritchie92 (talk) 16:40, 20 April 2020 (UTC)

Re: Graph:Chart: makes sense.

Re the other item, if you mean the table-graph entitled "COVID-19 cases in Italy" at the top of section "Timeline", I see no percentages of daily test counts; there are percentages, but not of test counts. From looking again at the whole page, I find no graph showing cases are percent of tests. --Dan Polansky (talk) 06:09, 21 April 2020 (UTC)

You're right, I got confused. The percent graph (total confirmed / total number of tests, or daily confirmed / daily number of tests) is however a bit problematic: (1) the number of tests is not the number of people tested, but merely the number of actual tests (some people might have done the test twice, three times, etc, because of checks, and this goes in the count of the number of tests); (2) the daily number of tests refers to the samples taken on that day, not to the samples analyzed on that day (i.e. the result of a test might be given on the same day but also the day after depending on the lab), so should the ratio be done as (no. of new cases on day "x" / no. of new tests on day "x") or (no. of new cases on day "x" / no. of new tests on day "x-1")? And analogously with the total cumulants: do we use the full cumulant, or do we subtract the latest daily number of tests? --Ritchie92 (talk) 07:49, 21 April 2020 (UTC)

UPDATE: now the Protezione Civile is starting to give also the number of people tested, starting from yesterday. Yesterday's numbers are still provisional, as analyzed in this article. This would solve the issue number (1) – starting from 23 April – that I explained in the previous paragraph. --Ritchie92 (talk) 08:46, 24 April 2020 (UTC)

Ritchie92 I am highly in favor of this version. Not only is it easier to comprehend for the readers but also removes the excess clutter of dates which are unpleasing to the eye, confusing and overall look messy as the dates increase. Editors familiar with graphs know what I mean, there will be a point when there's not even an inch of space left between dates and they all look like jumbled numbers. Only a slider would fix that issue but this parameter of x type=date outweighs the slider in terms of neatness and understanding. After all, the point here is for the readers to get the information as easily as possible. Personally, I don't see any logic behind this dispute. Canada has adopted this parameter since the start and the United States article also employs this.

Another issue I think we must start discussing are the bar graphs. Again, the same clutter of dates plus increasing width of these is problematic. If we employ the same parameter of x type and remove type=rect, this gives us a far cleaner version of the graph. Like below, albeit without the values, which honestly, are already displayed in the tables above, so are they really needed?

--Shawnqual (talk) 00:10, 25 April 2020 (UTC)

@Shawnqual: I totally agree with you (your version is very similar to what I edited a few days ago, and triggered this discussion). I would actually even remove the dates until 21 February, because they are not very significant in the monitoring of the pandemic development, and would reduce the size of the image significantly. I don't think the numbers on the bar charts are that important, but as soon as I removed them we got some complaints on this talk page, but of course as you say when the dates will be very close to each other the numbers on the bar chart will be unreadable, unless we do a slideshow which in my opinion is also unreadable. But regarding the other graphs (the ones that are currently not using the bar chart) I would switch to the date format as soon as possible (at the moment the graphs are already 900px wide).

@Nick.mon: what do you think? --Ritchie92 (talk) 08:05, 26 April 2020 (UTC)

Well, I would keep the current version regarding the daily numbers of deaths/cases/recoveries because I think a version where the daily numbers are showed, is useful and clearer, even if the reader has to scroll the graph. Regarding the first graph instead, I think that we can use the “short” version that you proposed. Anyway, this is only my opinion, I can live with your version too (all the more so if I don't have the necessary consensus) but I simply think that the one with bars would be better. -- Nick.mon (talk) 08:14, 26 April 2020 (UTC)

Alright then, it's settled. We use the date parameter for all graphs but the ones with bars. @Nick.mon:, while I do understands your perspective, I think you will likely change your mind about bar graphs as well later on. The issue would be this, even after we add the scroll for bar graphs, the width will continue to increase as dates roll by, (unfortunately, as we do not know when all this will end), so the default when a reader comes to the article will be perhaps the first month and a half displayed and without the legend too. This prevents readers from getting an overall view of the statistics. Anyways, we can revisit this discussion in the coming days, weeks. -- Shawnqual (talk) 08:40, 26 April 2020 (UTC)

If and when we decide to switch to the date type also for the daily graphs, I would say we can use the points for these ones, and use just the plain lines for the cumulative ones, because the cumulant is the one that is closer to a line, while the daily data are more scattered (so a point-wise plot would look better). --Ritchie92 (talk) 08:58, 26 April 2020 (UTC)

Yes, probably I was wrong, maybe having the same scale for the two types of graph would be better. Anyway, should we apply these improvements to the other articles (which, according to me, is the most important thing)? -- Nick.mon (talk) 09:07, 26 April 2020 (UTC)

@Ritchie92:, using points for the daily ones could work as well. But using plain lines for the cumulative ones hinders the impact of the information and progression scale as well. Canada's were only lines before I edited them, they lacked the aesthetic impact in comparison as all that could be seen was increasing plain lines. You have made one of those above, have a detailed look on it and also at the point ones right above, which do you prefer? The points also represent days, showing the increase in data, this information is lost with lines. If lines are used for daily graphs, the appeal and impact is different as there are fluctuations in data, showcasing the rise and falls of it, check the one I made above. In a nutshell, graphs with only rises (cumulative ones) aren't fit for lines in my opinion. I also think @Nick.mon: has a point for using same kind for both types of graphs, then I would propose, using points.

While I have your attention, I am thinking that the statistics sub-section looks quite jumbled up and is not easier to navigate. To solve this, I suggest dividing it into two sections of National and Regional. To do this, the source-code for charts of Lethality by gender and age could be written under the National sub-section, as they do not pertain to regions which the above two charts on the Template:2019–20 coronavirus pandemic data/Italy medical cases do but rather have the data for the whole country. Second, the Daily COVID-19 cases in Italy by region table could be collapsed on the main article. The national and regional graphs which currently are all together could be moved to their respective sections, this would make the section neater, easier to comprehend and navigate in my opinion. An example of what I mean is, here -note how the timeline table is collapsed at the very end, and here -the timeline table has been added as See also at the top. What do you think? ---- Shawnqual (talk) 09:28, 26 April 2020 (UTC)

In a nutshell, graphs with only rises (cumulative ones) aren't fit for lines in my opinion. I have the opposite opinions: graphs which have a more smooth behavior (cumulative ones) are better with lines, because they are anyway more similar to a line than to a scattered plot, whereas graphs which have a less constant behaviour (daily ones) are better with scattered points (I would even delete the lines joining the points).

I think it is better to keep all tables inside the article Template:2019–20 coronavirus pandemic data/Italy medical cases. Your proposal means that the data gets distributed between the template and tables written directly on this article, or in another template. I think this is more confusing than before. I don't think it's so much necessary to separate regional and national, to justify this big reorganization and proliferation of tables in the various articles. Unless we put everything (including the graphs) in the template article, then we can reorder it as we please. (By the way I am attracted by the idea of making cumulative plots for each region as it's done in the Canada article, what do you think?) --Ritchie92 (talk) 09:48, 26 April 2020 (UTC)

I think we have to extend these edits to the other articles, because if this article remains the only one to use them, I think that, per consistency, we should restore the previous versions. As I said before, this is the most important thing to me, anyway I can be a bit biased, because, as you know, I prefer the version with bars. -- Nick.mon (talk) 16:11, 27 April 2020 (UTC)

This is a possibility (although making this change on all possible COVID articles means opening tens of discussions in all other articles), however I would like to stress that consistency between different articles is not a Wikipedia policy nor something that would justify reverting to a version that does not have consensus. So, the Italian article can perfectly have a graph with a different format than the Canadian and the Spanish one, there is no problem with this. --Ritchie92 (talk) 16:24, 27 April 2020 (UTC)

Yes, you're right, it would be quite impossible, maybe we should start a discussion on the main article. Anyway, probably I'm only "desperately" trying to restore the previous version :) because I don't see so many differences between this version and your one, regarding the time scale and the width of the graph. -- Nick.mon (talk) 16:32, 27 April 2020 (UTC)

Of course the width now is sort of acceptable. In a few weeks it will look like this:

Or it will be larger and larger:

--Ritchie92 (talk) 18:54, 27 April 2020 (UTC)

Yes, but even the current version, right? -- Nick.mon (talk) 20:19, 27 April 2020 (UTC)

No the current version will not show all the dates on the x axis, only one every week, so it can be "squeezed" a lot more without the problems of having the dates text going one over the other. Like this, more or less:

--Ritchie92 (talk) 21:11, 27 April 2020 (UTC)

Ok ok, I understand, anyway I still think that a scrollable graph with bars, which highlights the daily number of cases/deaths/recoveries is better than a linear graph without this detail. Anyway, it seems that I'm the only one who thinks that. -- Nick.mon (talk) 13:13, 28 April 2020 (UTC)

MD-wiki-user, I saw your edits and I'd like to inform you that there's an ongoing discussion here, regarding the graphs, if you want to contribute. -- Nick.mon (talk) 21:26, 3 May 2020 (UTC)

@Nick.mon: @Ritchie92: I also believe that a scrollable graph with data shown for each day is very useful because everyone can seeexactly the daily variation, and at the same time the evolution curve can be observed. Additionally, it should be taken into consideration that now most people are using smartphones or other mobile devices to search and read info on-line, so they are seeing articles (including on Wikipedia) holding their devices in vertical position, and therefore, the scroll of charts is inevitable. Even current charts in the article "2020 coronavirus pandemic in Italy", when seen from a smartphone, need to be scrolled.

So, assuming that the pandemic will not end totally in the next months (which is highly probable), I propose that for the next 1-2 weeks or so to display a scrollable chart, but with an acceptable width, not very wide, not very narrow; to ensure readability, the "xAxisAngle=" can be changed from "-60" to "-90". But for the long term, as a compromise, but also to ensure more transparency, I believe that 2 types of charts could be presented. The first, on top, could be a line chart, for the entire period, showing the general evolution. The second chart, right bellow, could be a scrollable bar chart, with data for each day. Such a solution will satisfy all types of readers: those who need to see only the general evolution, without detail; and those who are interested to see more, and are not getting tired to scroll 1-2 sec for this (I know that the existing table is already showing the daily data, but I think we all agree that a picture/chart is more powerful that any text or table). Thus, I propose something like this:

No. of new confirmed cases per day

--MD-wiki-user (talk) 23:12, 3 May 2020 (UTC)

We already have too many charts. Duplication is not an option. I might be crazy, but I think that looking at a chart without having to scroll through it until the end is a huge advantage (and remember that this pandemic can last a year or so!) and it means that one can get the full picture of the pandemic behaviour in one look. Understanding trends and long-term variations on a graph that is three pagewidths long is not easy. So I am strongly against the proposal of having a scrollable graph just for the reason of having the numbers on the plot. The numbers are already in the tables! Also, we cannot tailor our layout uniquely for the smartphone users. --Ritchie92 (talk) 07:23, 4 May 2020 (UTC)

Could the number labels for the y-axis be both at the right and at the left rather than only at the left? That would ease seeing the deaths counts for the rightmost part of the graph.

I find bars better than the line; in future, they can be made very thin to accomodate more data onto one screen without scrolling, and then the number labels on the bars can be dropped since they would no longer fit it, and we would rely on the y-axis labels only. --Dan Polansky (talk) 08:42, 4 May 2020 (UTC)

A)

B)

These are the two options. -- Nick.mon (talk) 08:53, 4 May 2020 (UTC)

I don't think these are the only two options: as Dan Polansky was saying we can have the bar chart with very slim bars and without the numbers on top (however this doesn't work with the type date so we would still have to adopt the scrolling area at some point). With respect to the choice between (A) and (B) I am strongly against B and in favour of A, as it might be inferred. I find it unnecessarily uncomfortable and extremely unclear to have such a long scroll, and losing sight of the general trend (it even makes the presence of the plot meaningless, in my opinion). --Ritchie92 (talk) 09:01, 4 May 2020 (UTC)

Ritchie, but if this pandemic will last a year, as you said, (and unfortunately, I think you're right) we'll have to use the scrollable version for A too, and at this point I think it would be better having a graph which shows the daily numbers. -- Nick.mon (talk) 09:09, 4 May 2020 (UTC)

Why don't we use the bar chart as long as it does not require scrolling, and only switch to another chart type as need arises? (I agree that scrolling is far from ideal and that there is a strong benefit in seeing an overview at a glance.) And for the thin-bar chart: is there a way to have a label only on, say, every 5th day on the x-axis? --Dan Polansky (talk) 09:15, 4 May 2020 (UTC)

@Nick.mon: I did not say that it will last a year, I said it can. And I don't see why keeping the numbers on top of the graph, if we might have a plot with 365 bars. At that point what's the need of these damn numbers? This is too much, and the perceived "advantage" of having the numbers on top of the bars is totally nullified by the extreme disadvantage of having a graph that is three, or five, or ten times larger than what it should be in order to be understandable and visible. The numbers are already in the tables above! Also, even though after a year we will have a graph that is too long we have other options in that case: for example we can split the graph into two periods of six months, or we can indeed decide to have a scroll, which will still be much smaller and browsable than the veeery long one that you are proposing, or anything else.

@Dan Polansky: If we go to bar chart at the moment it will already be at the borderline of "too crowded". I would stay within a maximum of 700-800 pixels of width, no more, otherwise all the long-time trends will look diluted. --Ritchie92 (talk) 09:22, 4 May 2020 (UTC)

And I didn't say that "this pandemic will last a year", I said "if this pandemic will last a year"... Ok, the daily numbers are already in the table above, but a graph is far more "attractive" than a table, and I think that having them in a graph would be quite better for a reader. -- Nick.mon (talk) 09:49, 4 May 2020 (UTC)

@Ritchie92:: Okay. What about a line graph with the area under the line filled? That would approach the bar chart a little. One thing that I like about the bar graph is that it emphasizes the area under the curve. --Dan Polansky (talk) 09:57, 4 May 2020 (UTC)

having them in a graph would be quite better for a reader, it would be definitely quite worse for the reader having to scroll ten times to see the plot from start to end.

@Dan Polansky: that's a possible alternative, however it would go a bit against the meaning of the graph (showing the number of daily cases as a discrete variable, not as a continuum), but anyway it would look like this:

Ritchie92, thank you, that looks pretty fine. The continuum effect is a consequence of using the line graph, I think; only the bar chart properly suggests discrete days. Maybe adding circles at the data points like it was done in one of the graphs above would help. The fill suggests there can be integration/summation under the curve, and indeed, this is how we get the cumulative cases. I guess I still prefer thin-bar graph if its x-axis can be reasonably rendered. (I am not sure how much it matters what I prefer; you are doing the plotting.) --Dan Polansky (talk) 11:12, 4 May 2020 (UTC)

Yes, I like this proposal too! -- Nick.mon (talk) 11:22, 4 May 2020 (UTC)

Type area is incompatible with symbols (i.e. data points).

I have another proposal: merging all graphs this way: --Ritchie92 (talk) 12:46, 4 May 2020 (UTC)

New proposal

Looks fine. There is a benefit in having the daily new cases and deaths in one chart. I think daily recoveries are not so interesting but that would be for another discussion. What I am missing there are daily tests. --Dan Polansky (talk) 13:30, 4 May 2020 (UTC)

I really like this one, great job Ritchie! -- Nick.mon (talk) 13:41, 4 May 2020 (UTC)

Thanks. I am only in doubt about the fact that the deaths are squeezed vertically in this merged plot, so one cannot visualize them as if they had their own plot. I also would ask Nick.mon whether they can find better colors (note that in order to have overlapping colors one has to add "80" before the color code, so #FF6347 becomes #80FF6347 for example). --Ritchie92 (talk) 14:54, 4 May 2020 (UTC)

I used different shades of colors, but not much has changed... maybe we should use completely different colors, but I'd keep a consistency between the template and the graph. -- Nick.mon (talk) 15:08, 4 May 2020 (UTC)

Hey guys, might be a bit late to the party, but I would like to contribute my two cents. :) The issue I am finding with the above graph as Ritchie92 has clearly stated above is, it would go a bit against the meaning of the graph (showing the number of daily cases as a discrete variable, not as a continuum). Dan Polansky, you're mistaken, the continuum effect is not a consequence of using the line graph, we are also talking about the data of cases and not of days. Yes, only bar graph shows the data of every single day, but these other graphs give a direct hint at them to, by placing days at a regular interval. Naturally, there are only days between Mar 22 and Mar 29, and not months or years unless stated otherwise. You have also correctly stated that the fill suggests there can be integration/summation under the curve, and indeed, this is how we get the cumulative cases, but these are daily graphs and no cumulative ones.

Yes, merging the daily graphs is a good idea, but why not use the same type as other graphs (lines with dots as days) on this article for a) data comprehension in terms of progression being easier in them than this and b) consistency? I also think the height for this graph is unnecessarily too high, any reason for that? 350 seems to be fine, as I have done below. What I am proposing is this:

ALTERNATIVELY

Also, why leave the data of daily tests out if we're merging the daily graphs? Yes, I'm aware that doing so would make the data for the other three y-axes look minuscule, like this:

but, we can work around this by making a log graph, the height here has been computed and set according to the log cycles, so it would look like this:

The graphs with areas under them aren't easy to comprehend as the colors overtake each other, one can easily loose track. It also fails to give any information for a particular day on the x-axis. Consider this: How does one know which day in the peak is Mar 17, Mar 19..etc? Or should that not matter? For any reader, this information is pertinent and should be easily discernible. This precisely is why I was opposed to line graphs without dots when we first started discussing this issue of graphs, Ritchie92 may recall. They had instead proposed a scatter plot for the daily graphs (removing the lines between dots) back then. I am also with them and strongly opposed to sliding graphs for the various reasons clearly written above. For what it's worth, i will say again that bar graphs with sliders are extremely cumbersome and unpleasing for huge amounts of data like is the case for the pandemic. Nick.mon had also stated that it would be better to use the same type for all graphs, which I am on board with.

---Shawnqual (talk) 17:24, 5 May 2020 (UTC)

After having analyzed this kind of data for some days now, I've grown to like the plain line graphs with no area fill and no filled circles. However, I enriched the graphs with moving averages for smoothing, plotting the raw data in semi-transparent (lighter) colors and the moving averages in darked colors. An example for Sweden:

I've grown to like the simplicity and clarity. I now think it preferable over the filled area graphs and over the bar graphs. --Dan Polansky (talk) 13:37, 7 May 2020 (UTC)

Well, after all the discussion we had, Ritchie has every right to hate me, but I think the first graph that Shawnqual proposed, with lines and dots, is probably better than the current one. Anyway, I'd keep the daily tests' graph separated from the others. -- Nick.mon (talk) 16:36, 12 May 2020 (UTC)

Semi-protected edit request on 27 April 2020

Please change the section "Data quality issues" under STATISTICS :

Data quality issues Deaths statistics for Italy do not include the total amount of coronavirus victims who died outside hospitals, including dozens who died in different nursing homes across the country. Therefore, the official deaths statistics are considered an underestimate.

The author uses as references an article of a Newspaper who is not an official source.

The deaths statistics do include the victims who died outside hospitals, such as nursing homes. Here below the official stats, stating that on April 14th 6773 people died in nursing homes

https://www.open.online/2020/04/19/coronavirus-la-mappa-dei-contagi-tra-gli-anziani-in-trappola-nelle-case-di-riposo/ MauENG (talk) 11:59, 27 April 2020 (UTC)

 Not done: please provide reliable sources that support the change you want to be made. I don't think open.online is a reliable source. Aasim 05:07, 30 April 2020 (UTC)

This source says 48% of cases occurred in nursing homes. TGCP (talk) 09:20, 17 May 2020 (UTC)

A video on the Coronavirus spread

Hello

I don't know whether I am being appropriate, but I made a video on the Coronavirus spread in Italy from the 8th of March to the 8th of May weighting the cases on the population:

https://www.youtube.com/watch?v=A8UyjUHePJ0

Admins may consider including the video in the page, if considered approriate.

Farewell, Andrea — Preceding unsigned comment added by 94.39.201.19 (talk) 09:43, 23 May 2020 (UTC)

Major discovery regarding COVID-19 Epidemiology in Italy

Moved from Wikipedia Talk:WikiProject COVID-19 § Major discovery regarding COVID-19 Epidemiology in Italy

 – •Shawnqual• 📚 • 💭 21:48, 25 May 2020 (UTC)

A new study (preprint-no peer review) has been published on Medrxiv where they have analysed blood donor samples from Milan (hardest hit region of Italy from the beginning of the Italian Outbreak. Consider patient one in Italy was identified on February 20.

We considered 789 individuals donating blood between February 24th (the first week of the Italian outbreak) and April 8th 2020, whose plasma samples were stored for hemovigilance studies. At the start of the outbreak, the overall seroprevalence of SARS-CoV-2 was 4.6% (2.3 to 7.9; P<0.0001 vs. 120 historical controls). Conclusions: SARS-CoV-2 infection was already circulating in Milan at the outbreak start. Social distancing may have been more effective in younger individuals, and by the end of April 4.4-10.8% of healthy adults had evidence of seroconversion. Asymptomatic infection may affect lipid profile and blood count. https://www.medrxiv.org/content/10.1101/2020.05.11.20098442v1.full.pdf

At least 2.3% of blood donors had antibodies in their blood at the beginning of the outbreak. This is a game changer. Can we use this source or find better ones? It's been widely reported in Italy by major news sources but it is medrxiv. -- {{u|Gtoffoletto}}  ^talk 10:53, 24 May 2020 (UTC)

Preprints are not allowed as sources for Wikipedia medical articles, so no, we can't use it. If these apparent results make it to a peer-reviewed publication that complies with WP:MEDRS, we can use it, but not until then. Boing! said Zebedee (talk) 08:45, 25 May 2020 (UTC)

Should be published soon. I would keep our eyes open for similar studies coming out in general. Hemovigilance studies could be crucial in understanding the spread of the virus retroactively and figure out what the hell is really going on. -- {{u|Gtoffoletto}}  ^talk 09:18, 26 May 2020 (UTC)

Article needs updating

Timeline shouldn't end here with April 20th. That is almost two months ago. It's important to update this article with information what happened with Italy's reopening. I have no local knowledge, and don't speak Italian, but I am sure many wikipedians are up to this task. 96.248.77.91 (talk) 00:33, 17 June 2020 (UTC)

Evidence of CV in Italy already in mid-Dec19?

BBC reports the Italian National Institute of Health (ISS) saying they've found evidence of CV dating back to 18 Dec 2019, two months before the first confirmed case. -- DoubleGrazing (talk) 12:09, 19 June 2020 (UTC)

A Commons file used on this page or its Wikidata item has been nominated for speedy deletion

The following Wikimedia Commons file used on this page or its Wikidata item has been nominated for speedy deletion:

• Posto di blocco Casalecchio Aprile 2020.jpg

You can see the reason for deletion at the file description page linked above. —Community Tech bot (talk) 04:53, 28 June 2020 (UTC)

Blatant Proselytisation

"Samaritan's Purse" is mentioned only to promote their image, contrary to Wiki guidelines.

200.68.142.26 (talk) 06:22, 27 July 2020 (UTC) baden k.

I think that there is a wrong date

The last paragraph of the introductory sentence begins, "As of August 31..." but it isn't August 31 yet. — Preceding unsigned comment added by 2607:FEA8:C360:231F:A97A:3271:CD26:72C (talk) 20:39, 4 August 2020 (UTC)

How many Italians were unknowingly infected?

Between 1.5-6 million Italians, about 10% of the population at maximum, but they locked down as a country on March 10th, before the first of 16 German states (March 12), all of Spain (March 14), France (March 17) and the UK (March 23), these nations have estimated higher percentage or population of possible SARS-CoV-2 infections. 2605:E000:100D:C571:8C36:F847:196F:592 (talk) 00:37, 7 August 2020 (UTC)

All-cause deaths

Recently, I added the following to the article but got reverted; a discussion is at Talk:COVID-19 pandemic in the United Kingdom#All-cause death charts.

Weekly all-cause deaths in Italy based on mortality.org data, stmf.csv[1]:

mortality.org indicates the data for 2020 to be preliminary; above, the last two weeks available from mortality.org were excluded to prevent the worst effect of registration delay.

Weekly all-cause deaths in Italy for 0-14 year olds, based on mortality.org data, stmf.csv[2]:

mortality.org indicates the data for 2020 to be preliminary; above, the last two weeks available from mortality.org were excluded to prevent the worst effect of registration delay.

All-cause deaths in Italy in weeks 1-20, year by year, based on mortality.org data, stmf.csv[3]:

mortality.org indicates the data for 2020 to be preliminary; above, the last two weeks available from mortality.org were excluded to prevent the worst effect of registration delay. The above is not adjusted by population size.

--Dan Polansky (talk) 10:42, 12 August 2020 (UTC)

A Commons file used on this page or its Wikidata item has been nominated for deletion

The following Wikimedia Commons file used on this page or its Wikidata item has been nominated for deletion:

• Tbilisi TV Tower - March 23, Solidarity to Italy.jpg

Participate in the deletion discussion at the nomination page. —Community Tech bot (talk) 20:22, 12 November 2020 (UTC)

Request for comment - cases change metrics

Please take part in discussion here: Project COVID-19, Medical cases charts - change type — Preceding unsigned comment added by Kohraa Mondel (talk • contribs) 22:52, 5 December 2020 (UTC)

Evidence virus was in Italy by September 2019

Where do we put information like this?

It looks like the study should at least be mentioned here as well.— Vchimpanzee • talk • contributions • 18:23, 20 November 2020 (UTC)

It would be inappropriate to add that information to this article at this time. However, if you can find a source meeting WP:RS, you should definitely raise it here for potential addition to the article. --Yamla (talk) 18:27, 20 November 2020 (UTC)

The article I linked to had a link to this one but this was just a starting point.— Vchimpanzee • talk • contributions • 20:01, 20 November 2020 (UTC)

The World Economic Forum reported on it. Is that considered a reliable source? https://www.weforum.org/agenda/2020/11/coronavirus-italy-covid-19-pandemic-europe-date-antibodies-study Alec Gargett (talk) 02:24, 10 February 2021 (UTC)

Post-expand Include Size

This page was exceeding the Post-expand Include Size limit. I temporarily hid the medical cases chart since it was the single-largest contributor as a short term solution, but long term, the following things can be done to decrease the include size if there is a desire to re-add the chart:

• Summarize or otherwise reduce detail across the article. Many sections, such as "Foreign cases linked to Italy", are very long and overly detailed.
• Reduce redundant references. There are many sentences in the above section with multiple references where only one is needed (see Wikipedia:Citation overkill). There probably don't need to by almost 900 citations in this article.
• Move the detailed tables and graphs in the "Statistics" section to a different page (e.g. Covid-19 pandemic in Italy/statistics) and link to it (not transclude it). These tables and graphs are already very VERY long and take a while to scroll through.

Please preview the page before restoring the medical cases chart template to ensure that there is no error message at the top about the template size being too long. Happy editing! --Ahecht (TALK
PAGE) 13:23, 8 September 2020 (UTC)

not always only one reference is needed, the sentence might include aspects cited in all the in-line references. In any casy, the "Foreign cases linked to Italy" is probably the first section to be trimmed or simply put in a separate article. It's not so important for a reader who wants to know about the pandemic in Italy, IMHO. COVID-19 pandemic in mainland China does not care as much about this aspect, for example.--Alexmar983 (talk) 08:50, 26 December 2020 (UTC)

So the paragraph of the timeline was removed but not the final section of foreign cases. I am quite surprised...--Alexmar983 (talk) 19:18, 14 February 2021 (UTC)

After removing Template:COVID-19 pandemic, the post-expand include size is 2,082,456 – just inside the 2,097,152 byte limit. There is only room for a modest-sized navigation bar at the bottom, not a huge collapsed navigation book! wbm1058 (talk) 20:56, 19 March 2021 (UTC)