Vaccine by State/Department
Hi everyone!
In this visualization I represent the number of COVID-19 doses assigned to each "State"(Here we called them Department) within the Colombian territory.
Something strange that happened to me was, that the look up node did not recognize the strings in the column named Nom_Territory therefore, I had to introduce the departments' names manually with a Make String node. I guess, it is something related to the data format.
Cheers,
Juan
-
Vax_by_Dept.jpg
1.73 MB
- Vax_by_Dept.zip 8.1 KB
Vax_by_Dept.jpg
Keyboard shortcuts
Generic
| ? | Show this help |
|---|---|
| ESC | Blurs the current field |
Comment Form
| r | Focus the comment reply box |
|---|---|
| ^ + ↩ | Submit the comment |
You can use Command ⌘ instead of Control ^ on Mac
Support Staff 1 Posted by john on 09 Sep, 2021 01:31 AM
Juan,
Yes, there is something funny about the "Nom_Territorio" heading in your CSV file. It's only 14 characters, but when you feed it into a length node, NodeBox thinks there are 15 characters there. I looked for hidden spaces or special characters, but could not find what the problem was.
I have seen things like this happen before. Sometimes Excel introduces a hidden character into headings. You could try simply renaming that header to something else, maybe just "name", and see if that solves the issue.
If not, there is another workaround that works every time.
Feed your CSV node into a Keys node. This will produce a list with the two headings in your CSV. Feed that into a First node to pull out just the first heading.
Now feed the output of that First node into the Key port of your Lookup node. The Lookup node will then work correctly and return all the Department names.
By the way, in the hardcoded list of Department names you generated with the make_strings node, all but the first string has a leading space. Not terribly noticeable but can slightly affect label alignments and will cause problems if you ever need to do string comparisons. You can easily fix this by adding a Trim node, or just use my Key node trick to pull the names directly from the CSV.
I'm not sure converting the totals into need icon sizes is the best way to show the relative sizes. The small needles are very hard to see, especially against your patterned backdrop. A simple bar chart would be easier to see and compare. You could also consider grouping the smallest departments into an "Other" bucket.
If there is any correlation between the location of the departments and their relative totals, it might be more informative to show a map of the Columbian territory with each state colored on a gradient scale, maybe dark blue for departments with a relatively hight number of doses and light yellow for departments with few doses. You would need to add a color key showing the gradient or set of swatches showing intervening colors and dose range for each color.
It all depends what you want to emphasize, and what patterns you want your viewers to understand.
John
2 Posted by juan.rozo23 on 10 Sep, 2021 01:51 PM
Dear john, thank you very much for the troubleshooting, I tried your advice and it works wonders!
I decided for the small needles to make clear that these departments have received an extremely small amount of vaccines and attention from the authorities, even when these same departments account for more than 51% of the Colombian territory. Hence, in media these departments are vastly underrepresented and, ignored by the population in general therefore, putting them into the "Other" bucket would render them as invisible as always.
You are right about the backdrop background! It is a bit noisy. Next time I'll make the backdrop more subtle.
I have in my list of visualizations to do, one following the cartographic principles but, I haven't reach that one so far. I'll show it here as soon as I have it ready
I have already a bar chart for this data set(See image), that is why I wanted to try something different.
Cheers,
Juan
Support Staff 3 Posted by john on 11 Sep, 2021 03:55 AM
Juan,
I quite like your stylized bar chart. Clean and easy to understand.
You could make a similar viz with one needle bar for each department. Inside each needle you could do a stacked bar chart with colored segments for each supplier. You should probably normalize the needles based on population so that a full needle represents 100% doses for the adult population of that department.
This would result in a lot of nearly empty needles, but that might make for a more dramatic statement. Instead of arranging all 37 departments in one vertical list, you could arrange them in 3 columns. OR you could even center each needle based on the location of its department within a simplified map of Columbia.
Another interesting viz to try would be a bubble chart with pies at each bubble where
If you were extra clever you could scale the X and Y axes so the diagonal line X=Y would represent the average population density. That way bubbles for the more urban departments would appear above the diagonal line, and those for the more rural departments would appear below that line.
To pack in even more information you could then turn each sized bubble into a pie chart with the breakdown of wedges colored by supplier.
This is a tad complicated and might make a mess if most of the departments land in the same general X,Y positions and overlap. But it might reveal some interesting relationships. You should see small bubbles / pie charts for the rural departments and bigger bubbles for the more concentrated departments. The colors within the pie charts might point to regional biases for some suppliers.
So many possibilities! The nice thing about NodeBox is that it lets you play with wacky ideas like these, adjusting weighting factors in real time, to see what works and what doesn't.
This is a really interesting project. Thanks for sharing!
John