Sparely
DU

Data & Method

How the six baseline files become decisions, every data-quality fix applied, and every assumption — so each number can be traced and challenged.

Baseline data used

FileRole in the briefHow the model uses it
BoMSpare Part Master + Current Inventory PositionStock, static Min/Max, MRP type, Installed Qty per FuncLoc (installed-base population)
FuncLoc listEquipment criticalityA/B/C/S criticality (A highest) → part criticality, equipment class for failure-rate pooling
Notif HistoryFailure historyFailure rate per FuncLoc (installed-base prior), emergency frequency (downtime calibration)
Movement MaterialConsumption historyDemand events, recency weighting, abnormal-issue detection, order type split
PR POProcurement & lead timeUnit price, lead time mean/variability, supplier reliability, open POs (on order)
Mitem & MplanMaintenance planNext due dates of part-consuming jobs → known future demand

Planning date 01 Oct 2026; history 01 May 2025 – 30 Sept 2026. Database built 2026-11-04 14:10:20.

Pipeline

  1. 1. Clean & link

    Deduplicate BoM and PR PO, normalize 28 cycle formats, convert currencies, link Notif → Order → Movement and BoM → FuncLoc.

  2. 2. Split demand by cause

    Turnaround issues become event demand; preventive demand is replaced by the maintenance plan forecast; the rest is failure-driven.

  3. 3. Forecast irregular demand

    Recency-weighted compound Poisson per part, blended by credibility with an installed-base prior (failure rate of every FuncLoc the part sits on). Never-used parts get a data-estimated first-demand rate.

  4. 4. Add known future demand

    Part-consuming jobs from the maintenance plan, mapped to the parts they replace, at a usage factor calibrated on history.

  5. 5. ML early warning

    Gradient boosting trained on 12 monthly snapshots (features before the date, label = demand in the next 90 days), validated out-of-time, then refit and scored on the planning date.

  6. 6. Optimize each part

    Lead-time demand distribution (Poisson for slow movers, normal otherwise) → safety stock, reorder point, EOQ order quantity, Max. The service level is chosen to minimize holding + ordering + downtime-risk cost, or set by criticality.

  7. 7. Identical stock first

    Material numbers with the same description and unit are linked; a need is covered by a twin's stock above its own reorder point before anything is bought.

  8. 8. Recommend & route

    Reorder / transfer / reduce / monitor / maintain with reasons, health status, stockout and overstock risk; worklists per role, PR draft, decision impact, scenarios.

  9. 9. Prove by replay

    The policy is rebuilt from data before 1 May 2026, then the actual May–Sep 2026 issues are replayed day by day against today's Min–Max and the Sparely policy.

Data-quality checks and fixes

DatasetCheckResultAction
FuncLoc listMissing criticality33Set to S (lowest)
BoMDuplicate FuncLoc x Material rows230Removed (identical values)
BoMStock/Min/Max identical on every FuncLoc row of a materialTrueTreated as warehouse-level values; counted once per material
BoMNon Stock materials (min=max=0) still holding stock1683Flagged as uncontrolled stock in Inventory Health
Movement MaterialExact duplicate rows (same date, part, qty, order)863Removed (2.3% of issued units)
Movement MaterialMovement types presentIssueAll rows treated as consumption (goods issue)
Notif HistoryNotifications without an order553Kept; no part link
Notif HistoryNotifications in May 2025 (backlog spike, ~4x normal)1463Failure rates measured from 2025-06-01
Mitem & MplanDistinct cycle text formats28Normalized to days
Mitem & MplanUnparseable cycles0Excluded from schedule
PR POExact duplicate rows5420Removed; 6,189 unique lines kept
PR PONon-IDR lines51Converted to IDR with fixed assumed rates
PR POPO delivery before PR date / lead time over 365 days2 / 106Kept as records, excluded from lead-time statistics
PR PORequested delivery date before PR date324Kept; lead time uses PO delivery date
PR POOpen POs (delivery on/after as-of date)766Counted as on-order quantity in the inventory position
PR POSuppliers per material (max)1Single-source risk flagged
Movement MaterialIssue lines linked to a notification via Order16.0%Equipment link for the rest comes from BoM FuncLoc
Mitem & MplanABC Indicator equals FuncLoc criticalityTrueFuncLoc criticality used as the single criticality source
Mitem & MplanPlanned-task part usage factor (actual preventive issues / installed qty of due tasks)raw 0.716, 91/601 materials overlapFactor 0.716 applied (clipped 0.05-1.0)
Movement MaterialParts unused before the split that got a first demand after it267 of 5897 (split 2026-05-01)Sets the demand prior for never-used parts
ML early warningOut-of-time AUC / top-decile precision0.895 / 47.6% (base rate 7.1%)Trained on 59,848 part-months, tested on 14,962
Movement MaterialLast turnaround window (turnaround-type issue lines > 3x monthly median)2025-11-01 to 2026-01-31: 360 parts, 145,095 units; 36% of all units issued in 17 months fell in this windowUsed as the scope for the next-turnaround readiness check
ReplayPolicy inputs built from data before the replay window2026-05-01; 3,035 part-days of actual demand to replayBoth policies replayed day by day on the same demand and lead times
BoMMaterial numbers sharing an identical description and unit133 groups, 597 materials; 66 groups have one number at zero stock while another holds stockIdentical stock is used before buying; groups flagged for master-data review
PR POUnit price of short-notice PRs (requested <= 14 days) vs the part's median priceratio 1.000 over 1,021 linesNo rush premium in the data: default emergency premium 0%

Assumptions

AreaAssumptionValueNote
PlanningAs-of date2026-10-01Data ends 30 Sep 2026
PlanningForecast horizon365 days
DemandOrder type by order-number prefix201=Corrective; 202=Emergency; 203=Turnaround; 204=Preventive; 205=Other maintenance; 206=Other maintenance; 209=Other maintenanceNot confirmed by organizers
DemandTurnaround issuesExcluded from baseline, shown as event demand
DemandRecency half-life180 daysRecent demand weighs more
DemandInstalled-base prior calibration (kappa)0.4294Scales FuncLoc failure rates so total prior events match observed events
DemandCredibility constant2.0Events needed before history outweighs prior
DemandFirst-demand rate for never-used parts0.04092 per exposure-day267 of 5897 parts unused before 2026-05-01 were demanded after it
DemandFailure-rate window2025-06-01 to 2026-09-30May 2025 backlog spike excluded
Maintenance planPart-consuming taskslube oil change -> OIL; re-greasing -> GREASE; filter element replacement -> FILTER ELEMENT, SCREEN; final element overhaul -> DIAPHRAGM, GASKET, O-RING, PACKING SET, SEAT RING, SPRING, STUDBOLT, TRIM SET; pump overhaul -> BEARING, COUPLING, GASKET, IMPELLER, O-RING, RING, SEAL, SLEEVE, STUDBOLT; (compressor|turbine) major overhaul -> BEARING, BLADE, DIAPHRAGM, FILTER ELEMENT, GASKET, NOZZLE, O-RING, PACKING SET, RING, SEAL, SLEEVE, STUDBOLTInspection/calibration/test consume none
Maintenance planPlanned-task usage factor0.716Calibrated from preventive-order history
ProcurementPO-Item AmountUnit priceConfirmed by organizers
ProcurementLead timePO delivery date - PR dateBlended with category when < 3 POs
ProcurementExpedited (emergency) delivery time10th percentile lead time of the part's categoryFastest deliveries actually observed; = downtime days when a stockout stops production
ProcurementEmergency price premium0% (short-notice price ratio 1.000)Measured on 1,021 PO lines requested within 14 days
ProcurementFX rates to IDRUSD=16300; EUR=17700; SGD=12600; JPY=108
ProcurementPR type meaningPR01=Import / long-lead; PR03=Regular replenishment; PR04=Regular (other); PR05=Project / turnaround; PR06=Regular replenishment; PR09=Urgent; PR10=OtherNot confirmed
CriticalityRankingA > B > C > SConfirmed by the team; critical = A/B; a material takes the most critical FuncLoc it is installed on

Safety and governance rules

Company policy, not cost–risk scenarios: fixed rules applied on top of the cost model.

RuleDefinitionTodayEffect
Safety-critical element (SCE)Installed on a FuncLoc tagged PRV- or SDV- (safety / pressure-relief valve, shutdown valve)860 parts on 1,040 locations; 578 of these locations are rated SDowntime risk costed as class A, so 399 SCE parts are reorders today (Rp 8.9B); at least 1 spare always kept (1 part held only by this floor); never “potentially obsolete”; every reduction needs sign-off
Long-lead partCritical (A/B or SCE), 80th-percentile lead time over 90 days measured on its own POs (category estimates do not count)544 partsBadge and reason: the fastest delivery in its category is too slow for emergency buying, so plan ahead or hold a capital spare
Management sign-offBuys above Rp 1.0B; reductions above Rp 1.0B or of any SCE part22 buys and 985 reductions among today's recommendationsPending-approval list in the Management workspace; the PR draft and released capital wait for the sign-off
Decision logEvery approve, reject, undo and sign-off with role, time, quantity, value and reasonKept in the browser in this MVPAudit trail and CSV export in the Management workspace; a shared database in the pilot

The SCE tag rule stands in for the plant's SCE register, which replaces it in the pilot. 56% of the tagged valves carry the lowest equipment criticality, so criticality alone would under-protect them.

Master-data work list

Candidates for the SAP master-data team to check, not automatic changes.

IssueMaterial numbersStock valueSuggested SAP action
Same description and unit under several numbers597Rp 38.9BCheck interchangeability; keep one active number and link or block the others
Non Stock part with repeated or planned demand265Rp 69.1BChange MRP type to Stock Item and load the recommended Min / Max
Stock Item that the policy would buy on demand1,030Rp 1.54TChange MRP type to Non Stock; run the stock down
Non Stock part holding stock1,582Rp 2.12TStock outside Min–Max control: return, transfer or reclassify

133 groups share an identical description and unit; in 66 of them one number is empty while another holds stock (the Planner already transfers between them). The same description is not proof of the same specification, so each group needs a check before merging. Stock values use PO prices, or category medians for parts never bought.

Data gaps and the additional data we propose

Each gap limits a specific result; the proposal names the source and the business reason.

Gap in the baseline dataEffect todayProposed additional data
No part list per maintenance task (Mplan); 56% of plan items sit on a FuncLoc that has a BoMPlanned demand is inferred from the BoM with a calibrated usage factorSAP PM task lists / component lists per maintenance plan
83% of parts were never purchased in the data periodTheir value uses category-median prices (labelled “estimated”)Material master standard price / moving average price (SAP MM)
No production value per hour of downtimeDowntime cost per day is a scenario input; sensitivity shown aboveUnit production plan × margin per tonne (e.g. ethylene) per equipment system
Unmet demand and historical stock levels are not recordedStockouts in the replay are simulated from the policies' own starting levelsSAP stock history (MB5B) and reservation / backorder records
Every part has a single supplier in PR PO (124 suppliers)Single-source risk can be flagged but not re-sourcedApproved vendor list / source list (SAP ME01) per material
No link from equipment to its process system (only site-level nodes)Results cannot be split by process unitFuncLoc hierarchy (superior functional location)

Sensitivity of the decisions to the assumptions that cannot come from data

Production loss per downtime day, holding-cost rate and cost per purchase order are business inputs. How much do the recommendations move if they are wrong?

CaseReorderReduceParts changing actionAvg inventoryProduction stops / yr
Downtime cost × 0.51,0993,614335 (4.5%)Rp 210.6B0.6
Base case1,1603,505–Rp 260.0B0.31
Downtime cost × 21,2183,432284 (3.8%)Rp 303.4B0.2
Holding cost 15% / yr1,1723,471123 (1.6%)Rp 277.0B0.26
Holding cost 30% / yr1,1263,573194 (2.6%)Rp 231.3B0.47
Order cost × 21,1553,49923 (0.3%)Rp 261.6B0.32

Halving or doubling an assumption changes the action of only the share of parts shown; the direction of the recommendations is driven by the data (demand, lead time, price, criticality, observed emergencies).

Downtime calibration

Production loss per day is a business input (scenario page). How often a stockout actually stops production is learned from data: 132 Emergency notifications in 487 days = 98.9 per year, which today's static Min–Max must reproduce.

A
1.07%
B
0.53%
C
0.16%
S
0.05%

Suppliers (single-source exposure)

SupplierPO linesPartsMedian LTOn time
RAJAWALI INSTRUMEN INTERNASIONAL PT4745924 d39%
CAHAYA TEKNIK PERSADA PT3746241 d36%
CAHAYA INDUSTRI SENTOSA PT2902433 d42%
SAMUDRA SEJAHTERA INTERNASIONAL PT2753031 d41%
CAHAYA LOGAM PRATAMA PT2603058 d41%
KARYA ENERGI SENTOSA PT23210078 d38%
GEMILANG LOGAM MAKMUR PT1772227 d41%
GEMILANG TEKNIK LESTARI PT153632 d39%

SQLite database

data/caliber.db — read by every page and API

TableRows
assumption18
backtest3
consumption_by_type77
consumption_monthly5,910
data_quality25
demand_event12,835
forecast_monthly89,772
funcloc14,484
material7,481
material_funcloc20,423
material_twin3,614
meta34
ml_calibration7
ml_features7,481
ml_importance20
ml_metrics3
movement16,655
mplan_item10,153
notification5,795
planned_demand826
po_line6,189
replay_demand3,035
replay_material7,481
replay_planned944
supplier124
tam_scope360

JSON API: /api/summary · /api/recommendations · /api/materials/[id] · /api/forecast · /api/replay · /api/scenario (POST) · /api/pr-draft (POST) · /api/impact (POST) · /api/data-quality · /api/master-data