No issues but a question for you. How is co-occurrence data captured? Is it based on multiple ID’d individuals occurring in the same media asset? Is it two or more ID’d individuals seen in the same sighting/occurrence? Is it both?
If it’s both, that’s great. If not, could it be?
Also, I recall that the co-occurrence diagram was designed to be predictive - is the co-occurrence table also predictive? And is the table populated in the same way as the co-occurrence column in the list of encounters in a marked individual’s record?
Sorry, turns out I had a few questions on the topic…
How is co-occurrence data captured? Is it based on multiple ID’d individuals occurring in the same media asset? Is it two or more ID’d individuals seen in the same sighting/occurrence? Is it both?
An Encounter of both individuals must be in the same Sighting (a.k.a. “Occurrence”) to be displayed in the table or diagram and co-occur. Here’s an example:
When you submit a new photo on an Encounter, the Encounter should automatically be assigned to a Sighting and if two animals are detected in the same image, a duplicate Encounter should be spawned, automatically linking the two animals in the image by an Encounter. But it is through Encounters on the same Sighting that the database actually links them.
If it’s both, that’s great. If not, could it be?
See above. It should be done for you automatically, but it is not how the database actually asks the questions. That is via the Encounters on the same Sighting.
Also, I recall that the co-occurrence diagram was designed to be predictive - is the co-occurrence table also predictive?
Neither is predictive. They should both display co-occurrences defined through the linked Encounters under Sighting.
since we’re on the subject, I just deployed a fix we are testing on Flukebook to better make sure you co-occurrence graphs are display when co-occurrences are at low numbers. It should make a big difference in visualization on the individual page.
