I'll now run multiple parallel queries to gather all the data needed for the report. Let me fire off the first batch.

Let me try simpler queries to check connectivity first.

Good — sn_hosts uses description not name. Now let me fire all main report queries in parallel.

Good — 29 alerts current week, 13 prior. Now let me run the full detail queries in parallel.

Good data coming in. Let me now run the remaining queries for events, prior week patterns, zone comparison, and officer data.

I now have all the data needed. Let me compile the complete weekly enforcement report.

🅿️ City of Adelaide — Weekly Parking Enforcement Report

Reporting Period: Monday 23 March – Sunday 29 March 2026 Comparison Period: Monday 16 March – Sunday 22 March 2026 Generated: 30 March 2026 | Timezone: Australia/Adelaide (ACDT, UTC+10:30)

1. Executive Summary

The week of 23–29 March delivered a strong rebound in enforcement activity, more than doubling alert volume week-on-week, driven entirely by the Franklin Street 30-minute parking station.

• 29 overstay alerts detected this week — up +123.1% WoW (+16 alerts vs 13 prior week)
• 242 ANPR scan events captured — up +266.7% WoW (+176 events vs 66 prior)
• Single active enforcement point: Franklin Street 30-minute parking zone, Adelaide CBD
• 100% of alerts are Rule 102 (park longer than period allowed), all generated via automated ANPR (method 102-1800)
• 4 of 29 alerts (13.8%) have received an AI-review decision (all APPROVED); the remaining 25 are pending review as of report date
• Wednesday 25 March was the busiest day, accounting for 34.5% of weekly alerts (10 alerts, 100 events)
• Alert-to-event conversion rate declined to 12.0% (from 19.7% prior week), reflecting a higher scanning throughput week on which fewer vehicles exceeded the limit proportionally

2. Enforcement Activity Overview

Activity surged significantly this week, with both scan events and violation alerts well above the prior period.

Daily alert breakdown — current week:

• Sunday remains a consistent zero-activity day; no enforcement patrols or camera events recorded
• Saturday now active with 1 alert and 4 events — a new pattern not seen in the prior week (0 Sat/Sun activity)
• Wednesday's event spike (100 events) drove the lowest single-day conversion rate (10.0%), suggesting high vehicle turnover on that day

3. Alert Outcomes & Acceptance Rates

All alerts this week remain in pending/detected status as of the report date; outcome determination is in progress via AI and manual review.

AI-Assisted Review Pipeline:

• Prior week achieved 100% AI-review coverage with all 13 alerts APPROVED — a strong benchmark
• Current week AI pipeline is partially complete: 4 reviewed, 25 awaiting AI processing (likely queued; all reviewed alerts approved)
• Sample AI review remark (Alert for plate S608CTB, Wed 25 Mar): "Plate detected at 09:15 and again at 13:25 — approximately 4 hours 10 minutes later, well exceeding the 30-minute limit."
• No alerts have been manually accepted, rejected, or exempted this week — downstream officer review workload is building as alerts accumulate
• Formal acceptance rate will be calculable once review pipeline clears; based on AI signals, approval trajectory points to a high acceptance rate (100% AI-approval rate on reviewed alerts)

4. Zone Performance

All enforcement activity this week is concentrated in a single ANPR-monitored zone on Franklin Street.

Active Zone — Current Week:

• Zone pk_zone = 3333, linked to host pk_host = 22 ("Franklin Street 30 mins parking") and camera pk_camera = 28
• All alerts are uniform: single violation type (102), single method (102-1800 = 1,800-second / 30-minute time limit)
• Zone was active 6 of 7 days (Mon–Sat); Sunday coverage is absent

All-Time Top Zones by Alert Volume (context / benchmarking):

• Franklin Street 30-min zone is a recently activated ANPR location with 42 total alerts in the database (13 prior week + 29 current week), far below the high-volume historical zones
• 7,532 active zones exist in the system (of 10,322 total); only 1 generated alerts this week — representing 0.01% zone utilisation in the reporting period
• No zone-level acceptance rate comparison is possible this week as all alerts remain pending

5. Vehicle & Officer Performance

Enforcement is fully automated at this site; all activity is attributed to a single host platform with no officer-assigned alerts recorded.

Host / Enforcement Vehicle Performance:

• Single active host this week — a fixed ANPR camera installation (not a mobile patrol vehicle)
• Camera pk_camera = 28 is the sole data source; marked available = true in system
• No officer-attributed alerts detected (no informant-username or officer-id properties found); enforcement is fully ANPR-automated
• Alert method 102-1800 confirms automated 30-minute overstay detection with no manual officer input
• AI review acting as first-line reviewer: 4 alerts approved by AI to date; AI remark detail confirms multi-event cross-referencing and OCR validation (e.g., plate S583CWH confirmed across 5 events with consistent OCR reads)

Prior Week Comparison:

6. Trends & Patterns

Enforcement data shows a clear weekly rhythm driven by business-hours parking demand on Franklin Street.

6-Week Alert Volume Trend (from earliest available data):

• Alert volume trend: After a 4-week declining run (51 → 13), this week marks a sharp reversal; the first significant upswing in 6 weeks
• Event volume is the highest in 5 weeks (242), suggesting increased camera uptime or heavier traffic flow on Franklin Street
• Week of 16 Feb remains the all-time peak in dataset: 51 alerts and 816 events (2.7× this week's event volume)

Day-of-Week Patterns:

• Wednesday is the dominant violation day (34.5% of weekly alerts); potential link to mid-week CBD activity / market days
• Friday is consistently the #2 busiest day across both weeks (6 alerts each week)
• Saturday enforcement is emerging — 1 alert detected this week vs none in prior weeks; camera was operational (4 scan events logged)

Hourly Alert Distribution — Current Week:

• Three peak alert windows: 10:00–11:00 AM, 13:00, and 17:00 (5 alerts each for first two; 4 at 17:00)
• 11:00 AM is the highest-traffic scan hour (41 events) — peak turnover period
• 16:00 shows zero violations despite 18 scan events — possible clean turnover or driver awareness at end-of-business
• Operating window is strictly 08:00–17:00 — no camera events outside this range confirm a fixed schedule

7. Recommendations

The following data-driven suggestions aim to improve enforcement coverage, processing efficiency, and detection effectiveness.

Expand Zone Coverage: - Only 1 of 7,532 active zones generated alerts this week — investigate whether other high-historical-volume zones (Pulteney St Taxi, Gouger St Loading, Grenfell St Permit) have lapsed ANPR coverage or pending reactivation - Priority reactivation targets: Pulteney St (2,213 all-time alerts), Gouger St (1,939), Grenfell St Permit (1,737) — all historically >10× the output of Franklin Street's current pace

Clear the Review Backlog: - 25 of 29 current-week alerts (86.2%) have not completed AI review — expedite AI pipeline processing before alerts age - With prior week achieving 100% AI review coverage, a workflow or capacity issue has emerged this week; investigate AI review queue health - Establish a review SLA (e.g., all alerts AI-reviewed within 24 hours of detection) to keep processing current

Leverage Wednesday & Friday Patterns: - Wednesday accounted for 34.5% of weekly alerts — consider scheduling any manual officer patrols or secondary verification tasks on Wednesdays to supplement camera enforcement - Friday is consistently active (6 alerts both weeks); ensure camera is operating at full capacity on Fridays

Saturday Coverage Formalisation: - Saturday enforcement is emerging (1 alert, 4 events this week) — confirm whether the Saturday operating schedule is intentional and consistent; if so, document it formally and ensure AI review runs on weekends - Consider extending camera operational hours on Saturdays to match weekday 08:00–17:00 window

Improve Alert-to-Event Conversion Tracking: - Conversion rate dropped from 19.7% to 12.0% this week — while more events is positive (more scanning), the lower rate may indicate more compliant parking behaviour, or gap in detection logic - Investigate Wednesday specifically: 100 events but only 10 alerts (10.0% conversion) — disproportionate volume warrants camera calibration and rule threshold review

Establish Formal Acceptance Rate Baselines: - All historical alerts in this dataset appear to remain in "detected" (pending) status — no accepted/rejected/exported alerts are visible in any week - Action required: Confirm whether downstream alert processing (export to ticketing, officer review, and status updates) is flowing correctly; absence of any accepted/rejected alerts across 6+ weeks may indicate a pipeline disconnect

Report generated from live SenBOS PostgreSQL database. All times in Australia/Adelaide (ACDT). Alert outcomes reflect status as at time of report generation (30 March 2026). Acceptance rates will update as pending alerts are reviewed.