Insightful Kiwi

Measure

What the camera markers and shaded surface show — switch with the Measure buttons.

Relative activity (RAI) — detection rate for the chosen species group, effort-adjusted by camera-hours. Zoom out for a line-level RAI (the DOC protocol's per-line index); zoom in for the per-camera rate. The shaded surface interpolates it between cameras.

Priority (predator vs protected) — a composite where to act score: high (red) where the chosen predator is active and the chosen protected species is scarce (predator RAI high × protected RAI low). Exploratory — click a row in the table for the detections behind it.

Timing (predator ↔ protected) — per camera, the median time between a protected-species detection and the nearest predator detection. Red = they use the same ground close in time; the popup shows the real median gap. Exploratory.

Relative activity Priority (predator vs protected) Timing (predator ↔ protected)

Species

Predator(s)

Protected

Predator after only (stalking)

Download CSV

Camera deployment review

Highlight

OK Watch Concern Problem No deployment

Protected ↔ predator timing

For every protected -species detection on camera, this measures how long until the nearest predator detection at the same camera — the time gap between the two species using the same spot.

Two views: a distribution of those gaps (how often predator and protected are close in time vs far apart) and a seasonal trend of the median gap. Click a distribution bar to see the detections behind it, down to the record.

Rats are nearly everywhere, so the default predator is Mustelids — switch to the predator that matters to you.

Short gaps (left of the distribution) mean the two species share the same ground close in time; long gaps mean they mostly don't overlap.

The median gap by season rising over time ⇒ they're separating in time — what you'd hope to see where control is thinning predators or shifting their behaviour.
Falling / flat ⇒ predator and protected are overlapping as much as ever.
Predator after only restricts to gaps where the predator arrives after the protected animal — the “stalking” direction (following), rather than either order.

It's exploratory : sharing a camera close in time is suggestive of risk, not proof of a kill or even a meeting.

Pairing — each protected detection is matched to the nearest in time predator detection at the same camera ; the gap is the hours between them.
Direction — by default the nearest predator detection either side counts; Predator after only keeps just those where the predator came later.
Buckets — gaps are grouped < 1 h · 1–6 h · 6–24 h · 1–7 d · 1–4 wk · > 1 mo.
Seasonal median — within each season, the median gap (robust to a few extreme values), plotted over time.
Needs both on one camera — only cameras that saw both a protected and a predator contribute; a camera that saw only one never forms a pair.

Camera timestamps must be true clock times for this (trapping has no clock time, so it's camera-only). Treat proximity in time as co-occurrence risk, not a confirmed interaction.

For every protected-species detection on camera, how long until the nearest predator detection at the same camera. Short gaps mean they share the same ground close in time. Rats are everywhere, so pick the predator that matters.

Predator

Protected

Predator after only (stalking — predator follows the protected animal)

Gap distribution — click a bar for the detections behind it

Median gap by season — is the separation widening?

Duplicate window

Possible-duplicate window (minutes)

Gap distribution — monitoring targets vs the rest

Impact by species at this window

Inspect the images — is a flagged run one individual?

Camera

Species

Measure

How each trap is coloured and sized — switch with the Measure buttons.

Captures — total predators of the chosen group caught per trap in the selected period. Raw counts: catches are sparse, so a count reads more honestly than a rate at low effort.

Capture rate — captures per 100 trap-nights (effort-adjusted), so heavily- and lightly-checked traps compare fairly. Colour and size are clamped at the 95th percentile so one low-effort outlier (e.g. 1 catch in a few trap-nights) can't flatten the rest.

Servicing health — each trap judged as of the period end by its gap since last check — good / watch / neglected (including never checked); too few checks to judge a cadence → insufficient data . The buckets are calibrated to each reserve's own typical check cadence, not one global yardstick. Sized by days since checked (bigger = staler); dormant (6 mo+) and historic (12 mo+) traps are faint — toggle No records to see them.

Captures Capture rate Servicing health

Species

Predator(s)

Protected

Predator after only (stalking)

Download CSV

Trap check-frequency review

Is the trapping actually being run ? Two views:

By trapline — each trap graded on how often it's been checked this period: good / watch / neglected , so you can see at a glance which lines are slipping. The management view.
Over time — the mix of good/watch/neglected and the catch rate (per 100 trap-nights ) across seasons or years: is servicing improving, and is catching following it?

The By-trapline table is driven by the sidebar Period ; the Over-time trend deliberately ignores Period (it spans all of time). Click through for the traps behind a line.

Each trap's grade comes from its mean gap between checks — but “good” isn't a fixed number of days. It's relative to this project's own checking pattern, calibrated separately per reserve (a remote block and a roadside block have different normal cadences):

Good — checked at least as often as the reserve's typical (≤ its 50 th percentile).
Watch — slower than typical (≤ the 90 th percentile).
Neglected — slower than the 90 th percentile (the worst-serviced tail).

Two guardrails stop the percentiles branding someone unfairly when everyone's fine (or letting everyone off when they're not): a trap checked every 14 days or better is always good ; a gap longer than 60 days is always neglected .

Insufficient data — fewer than 2 checks in the period is too little to judge a cadence. Dormant (~ 6 mo+ since a check) and historic (~ 12 mo+) traps are shown faint — likely paused or retired, not neglected.

Cutoffs = percentiles ( 50 th / 90 th) of each reserve's own per-trap mean-interval distribution, computed at import and frozen on the cache — so they recalibrate when data is re-imported, not as you change the period. The resolved day values show in the table legend.
Guardrails — floor 14 d (always good) and ceiling 60 d (always neglected) override the percentiles.
Period-relative — a trap's gap is days since its last check up to the period end (no peek at later checks); the same trap ages to dormant then historic in later periods.
Catch rate (Over time) = catches ÷ trap-nights × 100 , so the trend is comparable across periods regardless of how much trapping was done.

Servicing buckets are per-reserve and frozen on import; the catch rate is computed live for the trend's seasons/years.

By trapline
Over time

Also show (otherwise active traps only)

Dormant (6 mo+) Historic (12 mo+)

Servicing over time

By season By year

Bait effectiveness

This ranks baits by how often they catch — captures per 100 trap-nights (a rate that accounts for how long each bait was out), or total captures if you switch Measure to “Total caught”.

Which bait gets the credit

A catch is found at a check, but it was caught by whatever was in the trap before that check — so each catch is credited to the prior check's bait, not the fresh bait set on the day it was found.

Group by: Full recipe vs Ingredient

Full recipe treats the whole bait set as one thing — “Ping Pong + Salted Rabbit” is its own line, separate from either alone. Cleanest to read, but it splits the data into many small recipes.

Ingredient credits every ingredient for the catches in intervals where it appeared, so one catch on a two-ingredient bait counts for both. Useful for spotting ingredients that show up in baits that work — but it is a correlation, not proof : if two ingredients almost always travel together, their numbers blur and you can't tell which one did the work. Read it as “appears in successful baits”, not “causes catches”.

What's left off

Baits with under 100 trap-nights of use, or fewer than the Min captures floor, are dropped — too little to compare (one lucky catch on a rarely-used bait would otherwise top the chart). The chart shows the top 15.

Digging in

Click a bar for the individual captures, click a catch to open its full record, or click a trap name to see that trap's whole check history and watch the bait change around each catch.

Captures by bait — attributed to the bait that was in the trap during the catching interval (the prior check's bait). Click a bar to see the captures behind it, then a trap to see its bait history.

Captures of

Group by

Full recipe Ingredient

Min captures

Measure

Capture rate Total caught

Trapping effectiveness — does checking more often catch more?

“Surely checking traps more often catches more?” In raw totals it usually looks that way — but two things can inflate that link even when how often you check makes no real difference. This screen strips both out so the data can answer, rather than assuming it.

Season. Catch rates rise and fall through the year — when the peak falls depends on the species and the place — and people often check more in the busier stretches. So checking and catching can move together for reasons that have nothing to do with cause and effect.
Effort. More checks simply mean more chances to find and record a catch. A raw count partly measures how much work was done, not how well the trapping worked.

To remove both, the chart uses a rate (catches per 100 trap-nights , not raw counts) and compares only within a season . What's left is the real question:

Per trap-night, and within a given season, does a trap that's checked more often actually catch more than one checked rarely?

It's a live question because most control here is kill traps : once one fires it can't catch again until someone resets it — so in principle a trap left a long time between checks spends much of that gap unable to catch. The chart lets you see whether that shows up in your data.

One panel per season . Within a panel, bars are grouped by how often a trap is checked — tightest (≤7 d) on the left to sparsest (>30 d) on the right ; bar height is the catch rate (per 100 trap-nights ); the figure above each band is how many trap-seasons it pools (across years).

Bars stepping down to the right ⇒ tighter checking catches more per trap-night — the hypothesis holds for that season.
Roughly flat ⇒ how often you check makes little difference; the link was season + effort, not checking.
Which season is busiest varies by species and place — switch species / reserve (top) to ask of one predator or area.

The two bars: Nominal vs Operational

Nominal counts every trap-night in full. A kill trap that fires (or is set off) early in a long gap still racks up trap-nights it had no chance of catching on — so under nominal effort, sparse checking reads as a low rate.

Operational strips out that dead time. Once a kill trap has fired it can't catch again until someone resets it at the next check, so the rest of that interval is dead : effort on paper, no chance of a catch. We don't know when in the interval it fired — taking the catch as equally likely at any moment, the trap was live for about half the interval on average, so an interval ending in a catch or sprung trap is credited as a half-night rather than a full one. (This is the standard Nelson–Clark sprung-trap correction — the expected live time under the simplest assumption about catch timing, not an arbitrary fudge.)

Operational therefore has the smaller denominator and the higher rate. If switching to it lifts the sparse (right-hand) bars and flattens the gradient, the apparent “checking more catches more” was really just traps sitting idle and dead between checks — not the checking itself.

Cadence = a trap's mean check interval over the data (the average number of days between its checks).
Catch rate = catches of the chosen species ÷ trap-nights × 100 .
Bands = ≤7 d · 7–14 d · 14–30 d · >30 d, by that mean interval.
Pooling = each (season-of-year × band) sums catches and trap-nights across all years of that season (e.g. every summer pooled, every autumn pooled), for sample size and to hold season-of-year fixed.
Operational trap-nights = nominal, but an interval ending in a catch/sprung event counts as a half-night.

Frozen on import — the bands recompute only when data is re-imported, so they don't shift as you click around.

Caveats

Small bands are noisy — bands under 100 trap-nights are dropped; the figure above each band is the trap-seasons it pools.
Which lines, not just how often — within-season holds season constant, but not which traplines are worked (keen volunteers tend to work the high-predator lines too), so read a gradient as suggestive, not proof .
Species mix — “All predators” is dominated by rats; pick a species (mustelids sit much lower) to see its own pattern.

Catch rate (catches per 100 trap-nights ) by how often traps are checked, within each season — so a seasonal catch peak isn't mistaken for a checking effect. Bars higher on the left (tighter checking) would mean checking more often catches more per trap-night.

Captures of

Reserve

Predators

Protected

Neighbourhood — protected, predators & nearby trapping

Pick one anchor — a single camera site , a monitoring line , or a whole reserve — and (for a site or line) a radius . The chart then tracks, season by season, three things in that neighbourhood:

Protected activity on camera (your chosen protected species).
Predator activity on camera (your chosen predators).
Predators caught in the traps in/around the anchor.

Click any point to see the records behind it. It's a focused, local companion to the Are we winning? network view — zoomed to one place.

Two questions, depending on the anchor:

Site / line — the local story : is a protected species holding on here, and what are the cameras and nearby traps seeing of predators around it?
Reserve — the alignment check : does camera predator activity move with what the traps catch? If cameras show lots of predator activity but traps catch little, the control may be thin or ineffective there; if both fall together, it may be working.

The two camera lines are rates (comparable across seasons); the trap line is a count of catches. Read them for shape and timing , not a one-to-one scale — they're different measures. Treat it as exploratory : it points you at places to look, not proof of cause.

Neighbourhood — the cameras and traps within the chosen radius of the anchor (a reserve uses all of its own cameras/traps, no radius), from the pre-computed spatial adjacency.
Camera rate — detections pooled over the neighbourhood's cameras ÷ their combined camera-hours, × 500 (the per-camera scale from config); net of likely duplicates.
Trap catches — predators of the chosen species caught at trap checks in the neighbourhood, as a per-season count .
Why trap is a count, not a timing — a kill is only known to the check window (we know it happened sometime between two checks, not the day), so a within- season count is the honest resolution; it can't be lined up to a camera detection in time.
Grain — toggle By season (every summer pooled, etc.) or By year (each season-year separately).

All exploratory and observational — neighbourhoods overlap and traplines aren't randomly placed, so read alignment as suggestive.

Pick an anchor — a single camera site , a monitoring line , or a whole reserve . For the cameras and traps in/around it, see season by season how protected and predator activity on camera move, alongside the predators caught in nearby traps. A trap catch is only known to its check window, so the trap line is a per-season count — not a timing.

Anchor

Site Line Reserve

Camera

Line

Reserve

Radius

Protected

Predator

By season By year

Click a point to see the records behind it.

Coverage — protected hotspots vs predator control

One map carrying both devices so you can see where the protected species are and whether predator control is reaching them. Colour says what a marker is; size says how much — one legend, no clashing colour scales.

The layers (toggle top-right)

Protected (camera) — green dots, size ∝ detection rate. The hotspots you're protecting.
Predators (camera) — red rings, size ∝ detection rate. Where predators roam. Off by default.
Catches (traps) — purple dots, size ∝ predators caught. Where removal is actually happening.
Traps — small grey dots, every trap deployed: the trapping field.
Boundary — the monitored footprint (convex hull of the devices) per reserve; hover it for the reserve name.

A green hotspot ringed by purple and grey is covered ; a green hotspot with little around it is a gap — somewhere a protected species is active but trapping isn't reaching.

Camera markers are a detection rate (per 500 camera-hours), so big vs small is comparable across cameras regardless of how long each ran. Catch markers are a count of predators removed. Period and reserve come from the sidebar; click any marker for its detail.

The Network density table below asks a prior question — is the network even dense enough to work? — and the Coverage gaps table ranks the gaps numerically.

Camera rate — per location, detections of the chosen species ÷ its camera-hours × 500 (the per-camera scale from config), net of likely duplicates.
Catches — predators of the chosen species caught at each trap in the period.
Boundary — the convex hull of a reserve's devices (its monitored footprint).
Marker size — scaled to the value and clamped at the 95th percentile, so one outlier can't flatten the rest.

All for the sidebar's period & reserve; everything redraws when you change them.

Where are the protected species, and is predator control reaching them? Green = protected on camera; red = predators on camera; purple = predators caught in traps; grey = traps. A green hotspot ringed by purple/grey is covered; one with little around it is a gap. Period & reserve from the sidebar.

Protected

Predator

Network density by reserve

Is the network even dense enough to work? Footprint area, traps & cameras per km², and the typical (nearest-neighbour) spacing — structural coverage, independent of any period.

Coverage gaps

Every monitoring line (its cameras) is ranked worst-first by how well predator control reaches it — turning the map's eyeball judgement into a list you can work down.

Each line gets a status (next tab) and the numbers behind it. Use it to find the protected hotspots where the trapping is thin, absent, or neglected.

Columns

Protected / Predator — detections on the line's cameras, a rate per 500 camera-hours (same units as the map). 0 = none seen on camera.
Traps — traps running within the gap radius (how much trapping reaches the line).
Caught — predators caught in those nearby traps this period.
Neglected — of the nearby active traps, how many are unserviced this period.
Traps/km² — trap density across the whole reserve, for context.

Status

No trapping — protected on camera, but no traps running nearby.
Predators uncaught — predators on camera nearby, but none being caught.
Trapping neglected — traps are nearby but mostly unserviced.
Covered — protected hotspot with serviced trapping reaching it.
No protected here — no protected species on camera on this line.

Neighbourhood — a line's “nearby” is the traps within the gap radius of its cameras. A bigger radius counts more distant traps as nearby, so more lines look covered — it drives the Traps, Caught and Neglected columns.
Rates — protected/predator are detections ÷ camera-hours × 500 (as on the map).
Neglected — from the trap servicing assessment (a nearby active trap unchecked this period); see the Trap review help for the cadence buckets.
Ranking — worst gap first (protected present, control weakest).

Why isn't the radius drawn on the map? It's applied per line — each line has its own circle around its own cameras, so a single circle would mislead. The table is where the radius lives.

Gap radius

Protected hotspots ranked worst-first by how well predator control reaches them, within the gap radius . No trapping = no traps running nearby; Predators uncaught = predators on camera but none caught nearby; Neglected = nearby traps mostly unserviced.