I'll fire off multiple parallel queries to gather all the data needed for the 7 report sections at once.

Good — the DB is responding. Now I'll fire all section queries in parallel.

Need to check the actual column names for zones and hosts tables, then gather more data in parallel.

Now I know the correct column names. Let me run the zone, host, officer, and coverage queries in parallel.

I now have all the data I need. Let me compile the full report.

All data collected. Here is the complete report:

📋 Weekly Enforcement Report — E2E Test Council

Reporting Period: 30 March – 5 April 2026 (Week 14) Generated: 6 April 2026 | Timezone: Australia/Adelaide (ACST/ACDT) Comparison Period: 23 March – 29 March 2026 (Week 13)

1. Executive Summary

The E2E Test Council enforcement system generated 30 parking alerts this week across a single monitored zone, a modest +3.4% week-over-week increase. All alerts remain in the initial pending state awaiting officer review or automated export. The AI-assisted review pipeline processed its first alert this week, returning a rejection due to vehicle identity continuity issues. A 5-week upward trend in weekly alert volume signals growing ANPR detection consistency.

• 30 alerts detected this week (+3.4% WoW vs 29 prior week)
• 143 ANPR scan events captured — avg 4.77 events per alert
• 1 active zone and 1 active enforcement vehicle out of 9 configured hosts
• 100% pending status — no alerts accepted, exported, or rejected this week
• 1 AI-assisted review completed → REJECTED (vehicle identity mismatch)
• 5-week upward trend: 13 → 17 → 28 → 29 → 30 — system stabilising
• No Sunday enforcement coverage recorded either week

2. Enforcement Activity Overview

This week's activity was generated by a single automated ANPR host over 6 active patrol days, operating exclusively within standard business hours.

• Total alerts: 30 (current week) vs 29 (prior week) — +3.4% WoW
• ANPR scan events: 143 linked to 30 alerts — 4.77 events/alert average
• Active patrol days: 6 of 7 (Mon 30 Mar – Sat 4 Apr); zero activity on Sun 5 Apr
• Operational window: 08:35–17:29 — enforcement confined to business hours
• Average alerts/active patrol day: 5.0 (vs 4.8 prior week)
• Alert type: 100% Type 102 — "Park longer than period allowed (30 min)" — no other offence types detected
• Alert method: 102-1800 — fully automated ANPR trigger after 1,800-second threshold
• Active generating hosts: 1 of 9 configured — 8 other hosts produced 0 alerts

Daily Activity Breakdown:

3. Alert Outcomes & Acceptance Rates

All 30 current-week alerts remain in detected (pending) status — no review, export, or disposition actions have been recorded for this reporting period.

Current Week Outcome Distribution:

• Acceptance rate: 0.0% — not applicable; no alerts have been reviewed yet
• Prior week status: identical — all 29 prior-week alerts also remain in detected status, now aged >7 days without action
• AI review activity (notable): 1 of 30 alerts (3.3%) received an automated AI review
• Decision: REJECTED
• Rationale recorded: Evidence showed multiple distinct vehicles occupying the space across the observation window; a red Kia Cerato (SA S052-CUL) was present ~08:00–08:08 and a white Land Rover Discovery Sport (SA XC-120N) appeared ~08:25–08:35; the OCR metadata plate (S941CGW) matched a background dark Audi — vehicle identity continuity for a single overstaying vehicle could not be established
• System-wide context for benchmarking: Across the full historical database, 50,233 alerts have reached ticket-exported status and 2,504 are exempted — confirming the review-to-export workflow is operational at scale in the platform

4. Zone Performance

All enforcement activity this week is concentrated in a single zone; no other zones generated any activity.

• Active zones this week: 1 — Franklin Street 30-min Parking (Zone key: 200208642_FranklinStreet_Adelaide_30minParking_1)
• Alert volume: 30 (CW) vs 29 (PW) — +3.4%
• Acceptance rate: 0.0% — all pending (no resolved outcomes to score)
• ANPR event coverage: all 143 events attributed to this single zone
• Zone consistency: sole active enforcement zone across all 7 recent active weeks
• Inactive zones: 0 alerts from any other zone in either period

Zone Summary Table:

• Hosts with zero zone coverage this week: Adelaide Prod, senforce-adelaide-01, senforce-adelaide-02, senforce-logan-01, adelaide-force-uat-01, COA Dev Server, Adelaide Revalidation — all active in system, none generating alerts

5. Vehicle & Officer Performance

A single automated enforcement vehicle accounts for all enforcement output; no human officer identifiers are recorded (fully automated workflow).

Host Performance Summary:

• Prior week (same host): 29 alerts over 6 active days — 4.8 alerts/day vs 5.0 this week
• Detection method: fully automated ANPR (102-1800) — no officer-name or officer-id properties present on any alert
• AI review system: active but low-throughput — only 1 of 30 alerts (3.3%) reviewed; returned 1 rejection
• Alert property coverage: 28 of 30 alerts (93.3%) have an alert-creation-timestamp property; 2 alerts (6.7%) are missing this property
• Idle configured hosts: 8 of 9 active hosts generated 0 alerts this week — significant untapped capacity in the fleet

6. Trends & Patterns

Enforcement activity follows a clear weekday-business-hours pattern with a triple daily peak and a strong multi-week recovery trend from a mid-March trough.

Day-of-Week Pattern:

Hourly Alert Distribution (Current Week):

Multi-Week Volume Trend (7 active weeks):

• Triple daily peak at 08:00, 13:00, and 17:00 — aligns with commuter arrival, lunch turnover, and end-of-work departure patterns
• No enforcement outside 08:00–17:00 — overnight and early-morning overstays go entirely undetected
• Wednesday high volatility: 10 alerts (PW) vs 4 alerts (CW) — largest single-day swing this week; warrants investigation for camera/connectivity issues
• Saturday persistently low: 1 alert each week for both periods — weekend parking likely lower-volume but pattern is very consistent
• 5-week recovery confirmed: alert volume has risen steadily from the 16-Mar trough of 13, now back at 30 — system is stabilising

7. Recommendations

The following actions are prioritised by urgency based on this week's data findings.

🔴 Immediate — Alert Review Backlog: - Investigate review pipeline: All 30 CW and all 29 PW alerts remain in detected status — prior-week alerts are now >7 days old with no action; determine whether the automated export trigger is misconfigured or inactive for this council - Establish review SLA: Define a maximum review-to-disposition window (e.g., 48 hours) to prevent staleness and potential statutory deadline breaches - Verify export configuration: System-wide data confirms 50,233 ticket-exported alerts exist historically — confirm this council's alerts are included in the export pipeline

🟡 High — Coverage Expansion: - Activate additional zones: 8 of 9 configured hosts are generating zero alerts; assess whether additional parking zones should be brought online to expand enforcement coverage across the council area - Address Sunday gap: Zero alerts on Sunday for both weeks — confirm whether the Franklin Street 30-min zone is restricted to weekdays; if not, weekend coverage should be activated - Extend operating hours: All detections fall within 08:00–17:00; no ANPR capture occurs outside this window — evaluate whether early-morning or late-evening overstays represent undetected violations

🟡 High — AI Review Pipeline: - Scale AI review to 100%: Only 1 of 30 alerts (3.3%) received AI review; the REJECTED outcome demonstrated real value in filtering a false positive — expanding AI pre-screening would significantly reduce manual review burden - OCR confidence thresholds: The AI rejection highlights a foreground/background vehicle plate mismatch; review OCR confidence settings to reduce false triggers from background vehicles in-frame

🟢 Medium — Data Quality & Monitoring: - Fix missing alert properties: 2 of 30 alerts (6.7%) lack the alert-creation-timestamp property — investigate whether the alert creation pipeline intermittently drops property writes - Wednesday variability investigation: Alert counts on Wednesdays swung from 10 (PW) to 4 (CW) with a truncated operational window (10:00–12:00 only) — check camera uptime and ANPR connectivity logs for Wednesday 1 April - Saturday patrol review: Saturday generates exactly 1 alert/week consistently; if the zone is monitored all day on Saturdays, consider whether patrol frequency or camera sensitivity warrants a review

Report generated from live SenBOS database queries — Adelaide timezone (ACST/ACDT). All figures reflect data as at 06 April 2026 02:00 ACST pipeline refresh.

Here's a summary of what was found and how each section was built:

Data highlights driving this report: - 30 alerts this week vs 29 last week (+3.4%) — all from a single fully-automated ANPR host monitoring Franklin Street 30-min parking - 100% pending (detected) — no alerts for either week have been reviewed or exported, representing a potential 7-day+ backlog for prior-week alerts - 143 ANPR scan events linked to the 30 alerts (4.77 events/alert) — confirming the camera is scanning at appropriate frequency - 1 AI review processed this week, resulting in a REJECTED decision — the AI correctly identified that OCR metadata matched a background vehicle (dark Audi, plate S941-CGW), not the foreground vehicle, and that no single vehicle could be confirmed as overstaying - Triple daily peak pattern at 08:00, 13:00, and 17:00 — consistent with commuter parking demand - 5-week upward trend (13→17→28→29→30) following a mid-March trough — system stabilising after the Feb-16 spike of 51 alerts - 8 of 9 configured hosts idle — significant untapped enforcement capacity in the fleet