A colony of simulated bats, one night at a time. Use it to try out ways of introducing the Louisville females to the Princeton colony and see how fast each newcomer becomes part of the colony.
1
Pick a scenario at the top.
2
Press Run. More seeds = more replicate colonies.
3
Colonies tab: watch the networks join. Play the slider.
4
Integration tab: how each newcomer settles in, week by week.
6
Compare tab: one dial, two values, same seeds, curves overlaid.
5
Change a dial on the left, rerun, compare. Hover a dial for what it means.
Every night
transfers, fasting rota
→
who sits together
→
grooming, bouts and returns
→
food to the fasted bat
→
bonds grow or fade
Dials worth turning
data our datasetsfit STRAND posteriorcalibrated tuned to Princetonliteraturebrief Bryson's specguess reasoning in the hover card
Reading the charts
one newcomer (seed 1)
median over all newcomers and seeds, 5 to 95% band
integration threshold
one newcomer in one seed
a transfer
six months
integration grooming a newcomer receives per resident around her, vs a typical resident. 1 = like a resident.
familiar share of her grooming, the part that goes to her own Louisville bats.
density pairs that groomed in the last 14 nights.
modularity Q 0 = origins mixed, 0.3+ = two cliques.
⚠ merge dials are guesses; replays of two real merges missed in opposite directions⚠ true process, not what a camera records⚠ if two designs differ by less than the band, they do not differ
run to start
newcomerresident femaleresident malegrooming, last 14 nightstransfersix months
These curves are from the last run; settings have changed since. Press Run to refresh.
How each newcomer settles in, week by week after her own arrival
Received grooming is relative to the median resident's grooming received from other residents in the same week.
All curves are the true process, not what a camera would record. Familiar share counts grooming to same-origin bats among everyone the mover groomed, in whichever enclosure she is.
Does it matter when a bat is moved? Week-8 outcome by order of arrival
Week 8 is used because "weeks to threshold" is cut off for late movers (they have less follow-up).
These curves are from the last run; settings have changed since. Press Run to refresh.
How connected each enclosure's network is over the whole run
How separated the two origins still are inside the Princeton enclosure
Schedule
Duration counts from the first transfer to the last move; the six-month mark (dotted on the timeline) is the line the plan should stay under.
Two values of one dial, everything else the same
Same scenario, same seeds (the box at the top), every other dial as currently set. Differences smaller than the bands are noise.
Outcome at week 8 and for the whole run (median over seeds)
Newcomers, week by week after arrival
Princeton enclosure over the run
Mechanism DAG: edges scale with their dial and vanish at zero
traitstatetreatmentbehaviour
Hover an arrow for half a second to read why the model has it and which dial scales it. Arrows thicken with their dial and vanish at zero.