I have an umbrella project that includes one collection project that far outshines the others. I’ve seen this before on other ones as well.
Suggestion: instead of what appears to be straight # of observations per unit length - change it to a log scale. At least some way that smaller project seem to shrink into insignificance.
Would also be good for the CNC and GSB projects.
The winner!!
And the also rans …
As someone who has worked in science education and communication, I have to say - understanding of log scales is very limited in the general public. I think this change would likely cause more confusion than benefit.
but - unlike a graph - we have an actual numerical value for each project.
Agreed.
Perhaps if there was a toggle on/off feature, so it could be used by the projects that want it? But I do not think that logarithmic should be the default, and I don’t think it would be useful for projects primarily intended for the general public such as the CNC (as opposed to projects which might be specifically targeting folks with a more scientific background/training).
The current display - with the winner totally overshadowing everyone else - is not a great welcome for newbies, with their much smaller projects
scroll down
hidden out of sight.
I can see 8 projects for local CNC then scroll DOWN. You need 1674 obs to even get a bar to show at all. Then - it would be fairer to ditch the bar and only show the numbers.
Why not have a log plot as a view option on all collection projects, so that anyone can look at the graphs with it if they want? Definitely not as the default, but having it as a clientside option seems sensible.
I agree that it makes sense to have it as an option. However there are all sorts of biases inherent in any presentation method. How about ranking as a proportion of community population? Or as an observation per square metre of land/water? Or as a proportion of estimated community biodiversity? Or whatever?
All graphs can have numbers displayed on them (or not). The project leaderboards are graphs with numbers. If the numbers are all that is important, we would just show them and no graph. The point of a graph is to communicate information more effectively than numbers alone. If a graph uses a scale that people don’t expect, it is likely to confuse them. If iNat showed this graph on a log scale, and then the raw numbers, it would be truly confusing, as the numbers would not be proportional to the graph. iNat would also need to add an axis and a legend noting the scale.
In general, log axes are not recommended for bar graphs; see https://www.graphpad.com/support/faq/graph-tip-dont-use-a-log-scale-on-a-bar-graph/ for some commentary.
For instance, in the example above, using a log scale will compress most of the other bars so much that it won’t be easily apparent how different they are. Using a log scale in this example would effectively lead to a loss of visual information and a less useful graph.
(I would be more convinced that the bar display is effective if they didn’t range down from gazillion to zero) But iNat has produced some wonderful graphs that make the info more available to us non-scientists.
I think you’re missing the point - I’m not suggesting that we change the appearance of the graph, only that we consider changing how the bar lengths are calculated. I don’t care how its done, log scale is just one way, like the richter scale, to minimize the difference between the most and least number of observations. No one has to see the actual log scale - its just a way making the presentation more useable. I can see a day when half of the collection projects have no bar at all!
Instead of say, 1" of bar for every 10k observations, every extra inch of displayed bar represents10x more observations.
I would argue that the current display is becoming less useful as time goes by as we get more extreme values.
To me - your initial screenshot - with all the technicoloured bars is visual clutter. Noise that hides the signal.
All I need is 3 columns
OK, but I wouldn’t need country and I’d retain the current ability to sort by observations, species or observers and maybe show all three values, not just the one being sorted on.
The bar graph really doesn’t add any useful info, especially since the numbers could be more prominent.
I would also like to sort by % at RG (CNC 25 still ranges from 4 to 58 !)
And I would keep identifiers.
Then choose which column to sort by. Up? or down?
%RG - good idea. Already calculated and available when you open the collection projects.
Probably not practical to modify the apps, but maybe just the website?
I only use the website ;~)
I mean to be fair, that’s not really an issue with the bars having log scale or not? There were over 600 cities for the CNC this year, I’m not sure there’s a way to display all of those cities without a significant amount of scrolling.
I think the screen space would be better used for a table with columns.
If you consider how ‘little’ info is displayed on that first screenshot.
I would not show %RG. This would just gamify RG even more. We already have large scale fabrication of observations in some CNC areas, likely motivated at least in part by the event’s gamification/competitive elements. If %RG is displayed, this would likely increase user motivation to add bulk agreeing (and often incorrect) IDs to increase %RG. We already see this in some other projects where users are encouraged/required to have RG observations. Adding %RG to leaderboards could seriously increase the problems resulting from CNC.
