Covid-19 Cases by Population and Health Region

[greenlineParticipant]

Greetings,

I just completed this, after a week of expert help from Pamela of VDT, and she asked me if I would post it here. I’m in no way a data expert, but I had some ideas about visualising Covid-19 data that I wanted to show, and DataGraph has come through admirably. But it wasn’t easy and some of the complications that Pamela solved for me I won’t be able to explain.

But that said, I’ll explain why I made this graph and why it could be useful, and a bit about how DataGraph achieved it (at least the parts that I do understand). Here’s the graph, covid-19 case data current to April 7:

Goals of this Graph:

1. Comparing by cases per million of population, rather than just totals (per country or region).

2. Including very high (Italy) and very low (Taiwan) countries.

3. Breaking down Canada (where I live) into provinces of interest, and then breaking down my province (BC) into its 5 government-mandated Health Regions (Fraser, Island, etc.)

4. My theory was that lumping the Health Region data into provinces and the province data into ‘Canada’, which other analyses usually do, was throwing away information that could be of use to people, who naturally want to know what’s happening around themselves.

How DataGraph achieved this:

1. The dataset available from Johns Hopkins University has the Country level and Canadian province level data, which I imported and flattened.

2. Making a column for “Location” and entering the populations of each country (and province) in another column beside it allowed an f(x) column to calculate the ‘cases per million’ for each.

3. The vertical dotted line at April 7 is made as “LastDate” and so moves itself whenever the JH dataset is re-imported to update the underlying data.

4. In the JH dataset there was a problem with missing days having been filled in with the same data (governments not announcing for a day; whover collected the data just filled in the same value in each day). This caused ugly (and inaccurate) flat steps in the curves. These were solved (thanks Pamela!) by using Delta and Interpolation columns.

5. All the above was repeated for the data from the BCCDC (BC Centre for Disease Control), where I found the data for the 5 BC regions. I made my own table manually from this, and imported it into DataGraph. I did my own fudging (interpolation) for the missing days, in advance, so that Delta/Interpolation step isn’t necessary for this dataset.

6. Both datasets can be shown on this same graph.

Result:

I believe this does achieve my aim of showing the people who live in “Vancouver Coastal Health” (where I live) where they stand, in Cases/Million of Population, relative to Canada, Italy, Taiwan, Quebec, and the other four BC Health regions. I also have close friends and family in Ontario, and it was easy to also make a version for them showing where Ontario stood. (Any country or province could be added or removed easily, since they’re all in the JH Global Cases dataset.)

Next I’m going to import the Global Deaths from JH, and continue with that, doing much the same.

Caveat: Overall, it’s been very difficult, and I have some concerns about DataGraph documentation. Pamela had to rescue me three separate times with Step-by-Step instructions. Without that it would have been impossible. :-0

Gl

[charles phillipsParticipant]

Looks good.

[Author]

Posts