i made many comments there, but maybe you’re talking about this:
this is created simply by starting a map in AGOL (or ArcMap) using a CSV as your source. instead of pointing to a CSV file on your machine, you would point to a CSV generated on the fly by the old (deprecated) iNat API. in that particular example map, this would be the CSV file or API request that you would reference: https://www.inaturalist.org/observations.csv?place_id=1&taxon_id=3.
as noted in the quoted block above, you can return only up to a maximum of 200 observations this way. you never answered my question about how many observations you expected to have in your project, but that could be a limiting factor with this particular approach. also, that particular API is considered deprecated. so i personally would not attempt to do new development using it, knowing deprecation means that it’s near its end of life.
in the same post, i also say:
in that tutorial, i provide an example where i make a map in AGOL and also a Story Map, albeit it’s using the “Classic” version of these tools, since the new versions of these tools were not yet released or had just been released at the time of writing.
you could combine the two approaches by doing the CSV option as a layer on top of the Tile option. this would allow you to visualize all observations and highlight the latest n observations, allowing the latest n observations to be clicked on and/or visualized in other ways.
if you want to do a fancier dashboard such as this: https://storymaps.arcgis.com/collections/ee2234e44bb74b0e92dc425a0c55b11d?item=1, and it needs work for more than 200 observations, you could do something similar to the CSV option i mentioned above, but you would have to provide the CSV data yourself and update it yourself. this means that you would store the CSV in a public repository such as AGOL or Google Sheets or some other cloud service. then you could make your maps, other components, and story map using the data in that CSV file. finally, you would have to set up some sort of process to update that data periodically. (how you do this is up to you and your skills.)
here’s an example of where i created a map in AGOL using a CSV stored in AGOL (second part of the post): https://forum.inaturalist.org/t/mapping-of-excel-data/30531/12. note that in that post, i mention that i had problems loading referencing the original source of the CSV data (from a government website), possibly because the file was too large or because the server was responding too slowly. these are the kinds of problems you’ll potentially run into when you go with this kind of approach. the dataset will at some point become too large / complex to handle in a web-based map interface where you’re trying to separately handle each feature. so in those kinds of circumstances, the best you can do is visualize via map tiles, as referenced earlier.
that’s the most i can say for now since you haven’t provided more details on exactly what you’re trying to achieve.