Handling False-positives #34
Mattk70
started this conversation in
Show and tell
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
I'll write this post to illustrate the workflow for tidying up BirdNET detections using Chirpity. Ultimately, you will end up with a clean archive of your records and the option to export a CSV that you can use as a basis for submissions to eBird, Trektellen, BirdTrack or whatever.
Here are the initial results of an analysis of one 3 hour file:
The first thing I'll do is scan the summary table for unlikely species in my region, and delete them. For example, Savi's Warbler is a pretty rare vagrant, so these detections can be discarded:
I'll repeat that process for any easily discarded species.
Now, I'll move on to species that are possible. When you select a species in the summary, Chirpity will jump to the first detection and highlight it. Pressing P will play that detection, so you can quickly ascertain if it's good. Tab will move to the next detection in the list. If the detection is wrong, you have two options:
The species dropdown has focus when the pop-up opens, so you can immediately type the initial letters of the species until you get the right one, or, you can pick from the drop-down. As you can see, it's also possible to add a call count, label or comments.
When I am happy with this list. I can export it to CSV (File > Export results to CSV). If I just want to export the results from one species, I can filter by that species before export, but if the view is unfiltered, a file with all the detections will be created. That file is here:
All_detections.csv
You can also export audio clips of your detections - optionally with a couple of seconds padding and/or a fade effect. However, the view must be filtered to do this, Chirpity can only batch export audio for one species at a time.
Beta Was this translation helpful? Give feedback.
All reactions