Last updated by Billie Goolsby, July 10, 2026.

These data comprise two Excel files, an R script, and sets of video (MP4) files needed to reproduce our analyses.

Conventions used throughout

Missing values. Empty cells should be read as not available / not recorded. Where a variable was intentionally not scored (for example, visitorsex in the camera-benchmarking sheets, which was not used in this study), the cell is left blank on purpose; those cases are noted in the relevant column descriptions below. To reduce ambiguity for reusers, empty cells may be filled with NA.

Units. Units for every variable are given in the descriptions below. In brief: times of day and durations are recorded as clock times (HH:MM:SS); durations are also given in seconds (s); counts are whole numbers; percentages are proportions between 0 and 1 (unless noted); and dates are YYYY-MM-DD.

Data_forpaper.xlsx

This workbook contains six sheets: CombinedVisits, FieldData, TadpoleVisitCounts, TadpolePoolCounts, TadpolePoolVisits_tadpoleonly, and CombinedVisitsEggsOnly_lessthan.

• CombinedVisits — all data from laboratory observations.
  • timein, timeout — the times of day a frog enters and exits the canister (clock time, HH:MM:SS).
  • timedifference — the difference between timeout and timein (HH:MM:SS).
  • numberofseconds — that time difference converted to seconds (s).
  • visitorsex — the sex of the frog that enters the canister (categorical).
  • location — which canister the parent visits (egg canister vs. tadpole canister vs. empty canister; categorical).
  • visittype — the action that occurred: egg care, tadpole transport, courtship, water hydration, or tadpole visit (categorical).
  • behavior — a broader grouping of visittype: egg, tad transport, mating, or pool visits (categorical).
  • locationscombined — the location of the visit: egg canister or water pools (categorical).
• FieldData — data from field observations.
• TadpoleVisitCounts — water-pool interactions used to compare how often parents visit empty pools versus pools occupied by tadpoles.
• TadpolePoolCounts — the summed counts and percentage effort parents devote to empty pools versus pools with tadpoles.
• TadpolePoolVisits_tadpoleonly — how often each sex visits a tadpole; comparable to FieldData.
• CombinedVisitsEggsOnly_lessthan — parents' visits to the egg canister on all days before day 12 (the day the father performs tadpole transport), used for a time-series analysis of parental time investment across tadpole development.

Perch_comparisons_cleaned.xlsx

This workbook benchmarks a Wyze v3 camera against a GardePro camera and summarizes treefrog perch use. It contains seven sheets: continuous, contiinuous_wyze, table, continuous_summedup, trigger, trigger_summedup, and triggercounts.

continuous

Behavior recorded during the 24/7 ("continuous") benchmarking of Wyze v3 against GardePro.

• date — date of recording (YYYY-MM-DD).
• timeinframe — time of day the frog enters the video frame (clock time, HH:MM:SS).
• timeoutframe — time of day the frog exits the video frame (clock time, HH:MM:SS).
• timedifference — the difference between timeoutframe and timeinframe (HH:MM:SS).
• timedifferenceinseconds — that difference converted to seconds (s).
• visitorsex — identity/sex of the frog in frame. Not used in this study; blank in most rows.
• camera — camera type used (wyze or gardepro; categorical).
• details — free-text commentary on the observation. Blank where no comment was recorded.
• secondsoff — number of seconds the GardePro recording is offset from the actual behavior (s). Blank where not applicable.
• missingrecordingsseconds — seconds of behavior ignored entirely by GardePro (s). Blank where not applicable.

contiinuous_wyze

Total seconds recorded by Wyze v3 in 24/7 mode, split into true versus false detections. (Note: the sheet name contains a typographical double "i.")

• seconds — raw number of seconds recorded (s).
• category — whether the recorded data reflect true or false behavior (positive/negative; categorical).
• camera — camera type used (categorical).

table

Treefrog perch/space use counts in laboratory settings; one row per frog.

• (frog ID) — individual frog identifier (categorical). This column currently has no header in the file; a label such as frog_id should be added.
• Use Floor — total number of times the frog used floor space (count).
• Use Walls — total number of times the frog climbed walls (count).
• Use Plants — total number of times the frog explored flora (count).
• Use Pool — total number of times the frog visited a water pool (count).

continuous_summedup

Summary of the per-event GardePro timing offsets from the continuous benchmarking.

• secondsoff — number of seconds GardePro was off from the actual behavior for each event (s). Please confirm this description matches your intended use of the sheet.

trigger

Behavior recorded during the motion-triggered ("trigger") benchmarking of Wyze v3 against GardePro.

• date — date of recording (YYYY-MM-DD).
• timeinframe — time of day the frog enters the video frame (clock time, HH:MM:SS).
• timeoutframe — time of day the frog exits the video frame (clock time, HH:MM:SS).
• timedifference — the difference between timeoutframe and timeinframe (HH:MM:SS).
• timedifferenceinseconds — that difference converted to seconds (s).
• visitorsex — identity/sex of the frog in frame. Not used in this study.
• camera — camera type used (categorical).
• details — free-text commentary. Blank where no comment was recorded.
• event — label pairing the same behavioral event detected by both Wyze v3 and GardePro (categorical).

trigger_summedup

Total summed recording time, in seconds, for the same behavioral event across cameras.

• TimeInSeconds — total seconds a camera recorded for a specific behavioral event (s).
• Camera — camera type (categorical).
• EventPair — the event being compared (categorical).

triggercounts

Total number of events recorded by Wyze v3 in trigger mode versus GardePro, split into true versus false detections.

• cameras — camera type (categorical).
• positivetriggers — number of accurate triggers corresponding to animal behavior (count).
• negativetriggers — number of false triggers not corresponding to animal behavior (count).
• posperc — positivetriggers as a proportion of all triggers (0–1).
• negperc — negativetriggers as a proportion of all triggers (0–1).

Video (MP4) files

Wyze videos are named by the day of the recording. Within each zipped file is a set of folders labeled 00-24 for the 24 hours of the day. In other words, 20250303.zip corresponds to the recording of March 3, 2025. Inside will be 24 folders for each of that day's hours, and each folder has 60 minutes of recordings. Note that trigger experiments will have less recordings, as recordings only occurred when the camera was tripped versus 24/7 recording. GardePRO videos are named sequentially with each of the trigger "DSCF00##". However, all recordings have the day and time of the trigger directly on each MP4 file for posterity. 

Number of events recorded by Wyze v3 in trigger mode versus GardePro, split into true versus false detections.

• 20250303.zip, 20250304.zip, 20250305.zip, 20250306.zip, 20250307.zip — Wyze videos saved per day in the 24/7 (continuous) experiment.
• GardePro_continuous.zip — GardePro videos from the 24/7 experiment.
• 20250819.zip, 20250820.zip, 20250821.zip, 20250822.zip — Wyze videos saved per day in the trigger experiment.
• GardePro_trigger.zip — GardePro videos from the trigger experiment.

R script

The R script reads the Excel data (with each sheet saved as its own Excel file) and reproduces the figures from our preprint using ggplot2. Our published figures look slightly different due to recoloring in Adobe Illustrator, but this step is not necessary to reproduce the analyses.

Sharing / Access information

• Preprint: https://doi.org/10.1101/2023.04.17.537238
• GitHub (camera setup): https://github.com/lobsu/WyzeSetUps
• protocols.io (improving camera focus): https://dx.doi.org/10.17504/protocols.io.dm6gp3p38vzp/v1