Bug: Some continents show 0% on scatter plot

rempsyc / busara_dashboard

The Missing Majority in Behavioural Science Dashboard

https://remi-theriault.com/dashboards/missing_majority

1 stars 0 forks source link

Bug: Some continents show 0% on scatter plot #47

Closed rempsyc closed 3 months ago

rempsyc commented 3 months ago

          the improvements make a huge difference! just a couple comments on this version:

the over-time tables and graphs look like they must be drawing on different versions of the data. for example, compare the graphs and tables below; in 2024 in the table, Africa is listed at 4.78%, but in the graph that percentage is 0

there's a similar inconsistency between the graphical data and the data listed in the waffle plots (compare the over-time graph to the waffle plot below; Africa is listed as contributing 2% of data in the waffle plot, but in the over-time graph, there is no time-point where it has above 0% of the data)

_Originally posted by @psforscher in https://github.com/rempsyc/busara_dashboard/issues/33#issuecomment-2120392742_

rempsyc commented 3 months ago

The waffle plot was correct and tables were correct, but the scatter plot for continents had a computation problem which has now been resolved. Thanks for reporting!

rempsyc commented 3 months ago

@psforscher in the end I detected some more discrepancies between the tables and waffle plots, and I have now fixed them. Imagine if we had removed the table in the beginning like some reviewers requested, fortunately we still have this data quality check 😅

Now though we have too many countries showing up (below)

Waffle plots are not made to be visualized with more than 6 or so groupings. So what I did is I only pick the first five countries and stacked all the rest in "Other". I hope that's fine.

I remembered that you were saying that the scatter plots by countries were also overwhelming because of the number of countries (and also colours). So I wonder if a similar strategy would be viable here. What if we only kept the first six or so countries in the scatter plot and merged the rest of the data as "other"? I'm not so sure about it but just sharing the thought with you while it is fresh.

psforscher commented 3 months ago

yes, i think adding "Other" makes sense, and i like the idea of doing this in the scatterplot too -- as long as the table still shows all countries, so that the data are accessible to those who want to know.

rempsyc commented 3 months ago

Ok. Then what's a meaningful number for the scatter plot? Top 6? Top 10?

psforscher commented 3 months ago

let's say top 10

rempsyc commented 3 months ago

With this change the scrolling bar disappears and the corresponding instructions are now obsolete 😅