What the AI writes back
Every analysis produces one data set and seven ready-to-use prompts — a full report, a formal notice, a plain-English update, a behaviour assessment, an evidence summary, a week-on-week comparison, and a text-only visual breakdown. Paste the same data into Copilot with a different prompt, and you get a completely different write-up. All seven below are built from the same underlying (fabricated) week of recordings.
The data every prompt starts from
A condensed excerpt of what gets pasted in — the real file also lists every monitored recording and every detected bark, often several thousand lines for a week of audio.
Overall Session
Recordings ran from the evening of Friday 14 March to Friday 21 March 2025, covering a full week of continuous monitoring. Of 980 audio files analysed, 312 contained at least one detected bark, with activity concentrated around early mornings and evenings rather than spread evenly across the day.
Total Barking
Frequency & Gaps
Barking was infrequent relative to the full monitored period, averaging one event every 5 minutes 26 seconds when active. The longest uninterrupted quiet stretch ran 11 hours 3 minutes, from 1:12am to 12:15pm on Monday 17 March — spanning files with no detections across that entire window, not just the gap between two events.
Barking Clusters
Events within 5 minutes of each other, grouped as a single cluster:
| Cluster | First bark | Last bark | Events | Bark time | Activity window |
|---|---|---|---|---|---|
| 1 | Fri 14 Mar, 10:14pm | Fri 14 Mar, 10:26pm | 34 | 41.2 sec | 12 min 33 sec |
| 2 | Sat 15 Mar, 6:02am | Sat 15 Mar, 6:09am | 21 | 25.6 sec | 7 min 32 sec |
| 3 | Mon 17 Mar, 10:45pm | Mon 17 Mar, 11:12pm | 58 | 1 min 47 sec | 27 min 16 sec |
| 4 | Wed 19 Mar, 5:18am | Wed 19 Mar, 5:24am | 17 | 19.8 sec | 5 min 25 sec |
| 5 | Fri 21 Mar, 9:03pm | Fri 21 Mar, 9:31pm | 46 | 1 min 12 sec | 28 min 43 sec |
Key Observations
- The longest single bark ran 38.4 seconds at 11:07pm on Monday 17 March, within Cluster 3.
- The most active recording was
REC-B_20250317_224502.wav, with 58 bark events in one file. - The longest quiet period — 11 hours 3 minutes — fell overnight into Monday 17 March, immediately before that evening's largest cluster.
- Most clusters are short and reactive (under 15 minutes), consistent with response to passing pedestrians, vehicles, or wildlife. Cluster 3's 27-minute span with sustained barking is the exception, and may warrant a follow-up recording to confirm whether it recurs.
Bark Density by Time of Day
Hour of day (0–23), bark events per hour.
Community Safety
[Date]Case reference:
[CASE-NUMBER][Owner Name][Property Address]
Dear [Owner Name],
Re: Recorded barking activity — [Property Address]
Following a complaint received by Council, audio monitoring equipment was placed near your property and recorded continuously from Friday 14 March to Friday 21 March 2025. This letter is to advise you of the findings and request that steps be taken to address the activity recorded.
Over the monitoring period, 1,847 individual barking events were detected, totalling approximately 58 minutes of barking across the full week. While this represents a small proportion of the monitored time overall, several concentrated periods of activity were recorded, the most significant being a sustained episode of approximately 27 minutes on the evening of Monday 17 March, between 10:45pm and 11:12pm. Shorter episodes were also recorded in the early morning hours on several days.
Under [relevant provision of the Dog and Cat Management Act 1995 (SA) — to be confirmed with Council's compliance team], Council may take action where a dog's barking is found to unreasonably interfere with the peace and comfort of a neighbouring resident. We ask that you take reasonable steps within [14 days] of this notice to reduce the frequency and duration of barking, particularly during the overnight and early-morning periods identified above.
If you would like to discuss this matter or provide further information, please contact the officer below within the response period noted.
Yours sincerely,
[Officer Name]
Community Safety Officer
Alexandrina Council
[phone] · [email]
Thanks again for reaching out about the barking next door. We placed a monitoring device near the property and recorded audio continuously for a full week, from Friday 14 March to the following Friday.
Over that week, we picked up barking on a number of occasions — mostly short bursts early in the morning and in the evening — adding up to under an hour of actual barking across the entire week. One evening in particular, Monday 17 March, had a longer stretch of noisier barking lasting close to half an hour.
We've shared these findings with the property owner and asked them to take steps to reduce the noise, particularly overnight and in the early morning. We'll check back in a few weeks to see whether things have improved, and we'll be in touch either way to let you know the outcome.
Thanks for your patience while we worked through this — please don't hesitate to get in touch if the barking continues or gets worse in the meantime.
Likely pattern
Most detected barking falls into short bursts of under 15 minutes, concentrated around 5–6am and again from 9pm to midnight. This timing lines up with typical reactive triggers — passing pedestrians, delivery vehicles, other animals, or a household member's normal comings and goings — rather than prolonged distress.
The exception
Cluster 3, on the evening of Monday 17 March, stands apart: 58 events over 27 minutes, including the week's longest single bark at 38.4 seconds. A cluster of this length and intensity is more consistent with sustained anxiety-type barking than a reactive burst. Worth checking whether this coincided with the owner being away from the property for an extended period that evening.
Welfare assessment
Barking accounted for only 0.58% of the monitored week, which on its own is unlikely to indicate a significant welfare concern. The one outlier episode is the exception worth following up on directly with the owner, rather than a sign of an ongoing problem.
Recommendations
- Ask the owner whether Monday evenings follow a different routine (e.g. a regular commitment that keeps them away longer than usual).
- Suggest enrichment (puzzle feeders, chew toys) for early-morning hours if the dog is left alone before sunrise.
- If the pattern seen in Cluster 3 recurs in follow-up monitoring, consider recommending a consultation with a dog behaviourist rather than treating it as a straightforward nuisance case.
- A second monitoring period isn't necessary yet — recommend revisiting only if a further complaint is received.
Evidence Summary
Monitoring period: Friday 14 March 2025, 8:00pm to Friday 21 March 2025, 8:40pm (168 hours 40 minutes).
Method: Continuous audio recording analysed using AI audio classification (Google YAMNet model) via the Dog Bark Detector tool.
Findings: 1,847 bark events detected across 312 of 980 recorded files, totalling 58 minutes 12 seconds of barking (0.58% of the monitored period). Average detection confidence 74%.
Three most significant incidents:
- Monday 17 March, 10:45pm–11:12pm — 58 events over 27 minutes 16 seconds, including the longest single bark recorded (38.4 seconds).
- Friday 21 March, 9:03pm–9:31pm — 46 events over 28 minutes 43 seconds.
- Friday 14 March, 10:14pm–10:26pm — 34 events over 12 minutes 33 seconds.
Assessment: [The data supports / does not support — to be determined by the assessing officer] a finding of unreasonable barking under the relevant provision of the Dog and Cat Management Act 1995 (SA), based on the frequency, duration, and clustering of events observed.
Recommended next steps: Issue a formal notice to the property owner; schedule a follow-up monitoring period approximately 4 weeks after notice is served to assess compliance.
Week 1 vs Week 2
Week 1 is the original monitoring period above. Week 2 is a fabricated four-week follow-up, run after a notice was issued to the property owner.
| Metric | Week 1 (14–21 Mar) | Week 2 (11–18 Apr) |
|---|---|---|
| Bark events | 1,847 | 612 |
| Total bark duration | 58m 12s | 19m 40s |
| Barking % of monitored time | 0.58% | 0.19% |
| Longest cluster | 27 min 16 sec | 6 min 05 sec |
| Longest quiet period | 11h 03m | 18h 40m |
Trend assessment
A clear improvement across every measure — bark events down roughly 67%, total duration down two-thirds, and no cluster in Week 2 approaching the length of the 27-minute episode seen in Week 1. This is consistent with the owner having taken action after the notice was issued.
Recommendation
No further escalation required. Suggest closing active monitoring on this case, with a note to re-open if a new complaint is received.
Time Block Activity (excerpt)
The full report covers every ~25-minute block across the week; shown below is a representative excerpt.
Barking Intensity Distribution
Duration Profile
Quiet vs Barking Proportion
Key Visual Observations
- Peak barking window: 9pm–midnight, with a secondary peak around 5–6am.
- Longest quiet period: 11 hours 3 minutes, overnight Sunday into Monday.
- Overall pattern is sporadic and reactive, with one clustered, sustained episode on Monday evening standing apart from the rest of the week.
- Activity is concentrated in two daily windows rather than spread evenly — consistent with responses to routine neighbourhood activity rather than continuous distress.
Questions, or something not working as expected?
Get in touch — issues, feature ideas, or general questions welcome.