How It Works

fledgling uses greedy submodular optimization to pick the set of hotspots that maximizes your expected target species.

For each hotspot it estimates a detection probability for every species - the chance you'd observe it on a single "average effort" visit during your date window. These probabilities are a Wilson score lower bound based on observation frequency on a rolling 7-day window centered on your selected dates across all years of historic data. This estimator is conservative and robust to small sample sizes.

This assumes detection probabilities are independent across hotspots and species, and that past observation frequency predicts future detection probability, both reasonable approximations, but not always true.

The optimizer works greedily: it first picks the hotspot with the highest expected target species count. Then, for each remaining hotspot, it updates the marginal gain, which is the additional species you'd expect after accounting for what you'd likely already see at the hotspots chosen so far.

Formally, the probability of observing species s across all selected hotspots is:

$$P(\text{observe } s) = 1 - \prod_{\text{hotspots } h}(1 - \hat{p}_{s \text{ lower, } h})$$

And the expected target species count is:

$$\sum_{s} P(\text{observe } s)$$

This is a conservative statistical estimate of the number of target species you are likely to encounter for your selected itinerary, but of course it will vary based on effort, weather, luck, and the whims of individual birds and populations at large.

Happy birding!

Nick

Got it

Welcome to fledgling!

fledgling uses eBird data to help you:

• Find hotspots with the highest detection probability for your target species
• Build and reorder a custom birding itinerary
• Explore results on an interactive map

It is perfect for planning birding trips in your backyard or anywhere else you might want to explore, whether you're doing a big year or just looking for that one nemesis!