BirdNET-Go – Three-Model Analytics Dashboard

[madriver]

I wanted to share a Home Assistant dashboard I've been working on to monitor BirdNET-Go's three-model inference pipeline.
I'm running BirdNET v2.4, BirdNET v3.0, and Perch v2 across five RTSP audio sources.
The dashboard uses BirdNET-Go's API to display pipeline health, processing performance, effective overlap, duty cycle, model comparisons, confidence distributions, and diagnostic activity.
The goal is to better understand how the three models are performing and how efficiently the inference pipeline is operating, rather than simply counting bird detections.
Credit goes to ChatGPT, which helped me design, develop, and refine the dashboard through quite a few iterations.
The image below is just a preview of the dashboard. Please open the attached PDF to see the complete dashboard, including the additional analytics, charts, and diagnostic information. The PDF is much easier to read than a screenshot.
Thought I'd share it with the community!

screencapture-gastonhome-madriveraccess-birdnet-v4-analytics-analysis-v3-2026-10-22-14_32_20.pdf