electricsheepafrica/warehouse-inventory-management

Warehouse & Inventory Management Dataset

Abstract

This dataset provides 30,000 simulated warehouse-level observations (10,000 per scenario) of health commodity storage, inventory management, and warehousing performance across three tiers of the pharmaceutical supply chain in sub-Saharan Africa. Each record represents one commodity category assessed at one warehouse during one monthly period. The dataset captures 40+ variables spanning warehouse infrastructure, storage conditions, inventory accuracy, FEFO compliance, order fulfilment, wastage (expiry + damage), capacity utilisation, temperature excursions, pest damage, theft, and downstream facility impact. Three scenarios: national CMS (82% inventory accuracy), regional warehouse (55%), district store (28%).

This dataset is entirely simulated. It must not be used for warehouse operations or procurement decisions.

1. Introduction

1.1 Warehouse Management in Health Supply Chains

Warehousing is the critical link between procurement and last-mile distribution. USAID GHSC-PSM has documented that effective warehouse management — including proper storage conditions, inventory accuracy, and FEFO (First Expiry, First Out) compliance — directly impacts commodity availability at health facilities.

1.2 Storage Conditions

WHO Good Distribution Practices (GDP) require controlled temperature, humidity, pest management, and security for pharmaceutical storage. However, UNICEF Supply Division assessments indicate that only 40-60% of SSA warehouses meet WHO GDP standards, with district-level stores frequently lacking basic infrastructure including temperature monitoring, generator backup, and pest control.

1.3 Inventory Accuracy and Wastage

Stock record discrepancies between physical counts and records are widespread, with inventory accuracy as low as 25-30% at district stores. Wastage from expired and damaged stock reaches 15-30% at sub-national levels, representing significant financial losses and contributing to downstream stockouts.

1.4 Rationale

This dataset integrates warehouse infrastructure, storage quality, inventory management performance, and downstream impact indicators for supply chain optimization research and warehouse management system development.

2. Methodology

2.1 Parameterization

Parameter	National CMS	Regional WH	District Store	Source
Inventory accuracy	82%	55%	28%	JSI/SIAPS assessments
Order fulfilment	78%	52%	30%	GHSC-PSM data
Wastage rate	8%	18%	30%	Warehouse audits
Storage adequate	75%	42%	15%	UNICEF assessments
FEFO compliance	70%	35%	10%	WHO GDP audits
Capacity utilisation	85%	65%	40%	Infrastructure data

2.2 Commodity Categories

12 categories: essential medicines, ARVs, vaccines (cold chain), laboratory reagents (cold chain), contraceptives, malaria commodities, IV fluids, PPE/IPC supplies, surgical supplies, nutrition commodities, medical device consumables, controlled substances (secure storage).

3. Schema

Column	Type	Description
warehouse_level	categorical	national_CMS / regional_warehouse / district_store
warehouse_size_sqm	int	Storage area in square metres
commodity_category	categorical	12 commodity categories
storage_requirement	categorical	ambient / cold_chain_2_8C / secure_ambient
criticality	categorical	critical / high / medium / low
has_WMS	binary	Warehouse management system
has_temperature_monitoring	binary	Temperature monitoring
has_generator_backup	binary	Backup power
storage_conditions_adequate	binary	Meets GDP standards
inventory_accuracy_pct	float	Physical vs record match
stock_record_up_to_date	binary	Records current
fefo_compliance	binary	FEFO practiced
order_fulfilment_rate_pct	float	Orders fulfilled completely
orders_backordered	int	Unfulfilled orders
wastage_rate_pct	float	Expired + damaged rate
expired_stock_value_usd	float	Value of expired stock
capacity_utilisation_pct	float	Space used
temperature_excursion_month	int	Cold chain breaks
pest_damage_reported	binary	Pest damage
theft_reported	binary	Theft/pilferage
inventory_issue	categorical	11 issue categories
stockout_at_warehouse	binary	Warehouse-level stockout
facilities_affected_by_stockout	int	Downstream facilities impacted

4. Validation

5. Usage

from datasets import load_dataset

dataset = load_dataset(
    "electricsheepafrica/warehouse-inventory-management",
    "regional_warehouse"
)
df = dataset["train"].to_pandas()

# Wastage by commodity category
print(df.groupby('commodity_category')['wastage_rate_pct'].mean().sort_values(ascending=False))

6. Limitations

• Simulated: Not from real WMS data or warehouse audits.
• No seasonal effects: Humidity/temperature seasonal variation not modelled.
• Simplified costing: Wastage costs are estimates, not actual financial records.

7. References

1. USAID GHSC-PSM. Warehouse management best practices.
2. WHO (2014). Good storage and distribution practices (GDP).
3. JSI/SIAPS. Strengthening pharmaceutical supply chains.
4. UNICEF Supply Division. Warehouse capacity assessments.

Citation

@dataset{esa_warehouse_inventory_2025,
  title   = {Warehouse and Inventory Management Dataset},
  author  = {{Electric Sheep Africa}},
  year    = {2025},
  publisher = {Hugging Face},
  url     = {https://huggingface.co/datasets/electricsheepafrica/warehouse-inventory-management},
  note    = {Simulated dataset. Not for warehouse operations or procurement decisions.}
}

License

CC-BY-4.0