Data Cleaning Summary
3,670 → 3,614
Production Rows
56 removed (69 anomalies, some overlap)
95 → 85
Safety Rows
10 removed (duplicates + future dates)
298 → ~270
Cost Rows
~28 removed/fixed
Production Data — Cleaning Log
| Tipe Anomali | Jumlah | Treatment |
|---|---|---|
| Missing values (kolom produksi kosong) | 15 | Hapus row — tidak bisa diestimasi tanpa data produksi |
| Negative production values | 7 | Konversi ke absolute value (data entry error — angka benar, tanda salah) |
| Duplicate rows (Date + WellID sama) | 20 | Hapus duplikat, keep first occurrence |
| Future dates (tahun 2026) | 5 | Hapus — di luar periode analisis 2025 |
| Zero production pada well "Online" | 10 | Flag sebagai anomali. 6 terjadi saat turnaround (valid), 4 di luar turnaround (data entry error → hapus) |
| Format tanggal DD/MM/YYYY | 12 | Konversi ke YYYY-MM-DD (standardisasi) |
Safety Data — Cleaning Log
| Tipe Anomali | Jumlah | Treatment |
|---|---|---|
| Missing values (Department/Severity kosong) | 5 | Isi Severity berdasarkan IncidentType mapping. Department yang kosong → "Unknown" |
| Duplicate rows (IncidentID sama) | 8 | Hapus duplikat, keep first |
| Inconsistent severity (tidak match incident type) | 4 | Koreksi berdasarkan mapping: Near Miss→Low, Medical Treatment→Medium, LTI→High, dll |
| Future dates (tahun 2026) | 3 | Hapus |
Cost Data — Cleaning Log
| Tipe Anomali | Jumlah | Treatment |
|---|---|---|
| Missing amount values | 8 | Isi dengan Budget_USD value (best estimate) |
| Negative cost values | 4 | Pisahkan sebagai credit notes. Untuk analisis, gunakan absolute value |
| Duplicate rows | 10 | Hapus duplikat, keep first |
| Typo kategori | 3 | Personel→Personnel, Maintanance→Maintenance, Logistic→Logistics |
Production Detail
Average Daily Production per Asset
| Asset | Avg Gross (boepd) | Avg Net (boepd) | WI | Avg Uptime (%) |
|---|---|---|---|---|
| Block A | ~13,000 | ~3,730 | 28.67% | ~96% (excl turnaround) |
| Tuna | ~800 | ~400 | 50.0% | ~96% |
| Total | ~13,800 | ~4,130 | — | — |
Well Performance Ranking (by Avg Gross boepd)
| Rank | Well ID | Asset | Platform | Avg Gross (boepd) |
|---|---|---|---|---|
| 1 | A-01 | Block A | Alpha | ~2,050 |
| 2 | A-03 | Block A | Alpha | ~2,010 |
| 3 | B-01 | Block A | Bravo | ~2,000 |
| 4 | A-02 | Block A | Alpha | ~1,980 |
| 5 | B-03 | Block A | Bravo | ~1,970 |
| 6 | B-02 | Block A | Bravo | ~1,960 |
| 7 | A-04 | Block A | Alpha | ~1,950 |
| 8 | T-01 | Tuna | Tuna-FPSO | ~280 |
| 9 | T-03 | Tuna | Tuna-FPSO | ~270 |
| 10 | T-02 | Tuna | Tuna-FPSO | ~260 |
Safety Detail
TRIR Formula
TRIR = (20 × 1,000,000) / 20,000,000 = 1.0
Recordable: Medical Treatment (10) + Restricted Work (4) + Lost Time Injury (6) = 20
Incident Breakdown by Type
| Incident Type | Count | % of Total | Recordable? |
|---|---|---|---|
| Near Miss | 30 | 35% | ❌ |
| First Aid | 16 | 19% | ❌ |
| Medical Treatment | 10 | 12% | ✅ |
| Equipment Failure | 7 | 8% | ❌ |
| Environmental Release | 6 | 7% | ❌ |
| Lost Time Injury | 6 | 7% | ✅ |
| Process Safety Event | 6 | 7% | ❌ |
| Restricted Work | 4 | 5% | ✅ |
| Total | 85 | 100% | 20 recordable |
Incident Breakdown by Asset
| Asset | Count | % of Total |
|---|---|---|
| Block A | 50 | 59% |
| Tuna | 23 | 27% |
| Andaman II | 12 | 14% |
Cost Detail
Unit OpEx per Asset
| Asset | Annual OpEx (USD) | Annual Production (boe) | Unit OpEx ($/boe) | vs Benchmark |
|---|---|---|---|---|
| Block A | ~$57.96M | ~4,745,000 | ~$12.2 | Below $12.8 |
| Tuna | ~$5.18M | ~292,000 | ~$18.0 | Above $12.8 |
Cost Breakdown by Category (Block A, Annual)
| Category | Annual Amount (USD) | % of Total | Budget Variance |
|---|---|---|---|
| Personnel | ~$17.4M | 30% | Within ±5% |
| Maintenance | ~$11.6M | 20% | +25% (Juni turnaround) |
| Logistics | ~$8.7M | 15% | Within ±5% |
| Marine & Aviation | ~$5.8M | 10% | Within ±5% |
| Chemicals & Consumables | ~$4.6M | 8% | Within ±5% |
| Insurance | ~$4.1M | 7% | On budget |
| Utilities | ~$2.9M | 5% | On budget |
| Other | ~$2.9M | 5% | Within ±5% |
| Total | ~$57.96M | 100% | YTD +3% vs budget |
⚠️ Disclaimer: Data operasional di-generate secara sintetis (random seed 42) untuk keperluan case study. Angka-angka di-anchor ke data publik Harbour Energy plc (LSE filings) tetapi tidak merepresentasikan data aktual perusahaan.