Las Taperitas S.A. is one of the largest and most important family-owned agribusiness companies in Argentina. Founded in 1928 in Santa Fe, the company operates across multiple business units — dairy and pork production, agriculture, cattle ranching, feedlots, and forestry — with presence in four provinces: Santa Fe, Neuquén, Entre Ríos, and Corrientes. Las Taperitas is also deeply connected to Ilolay, one of Argentina's most recognized dairy brands.
Managing a business of this scale and diversity generates an enormous volume of operational and financial data spread across multiple systems and areas. Accounting entries, purchase records, cost center allocations, agricultural cycles, currency conversions, and asset depreciation were all tracked in disconnected ways, making consolidated analysis difficult and time-consuming. Reporting relied heavily on manual processes, which limited both the speed and the reliability of the information available to management.
Las Taperitas engaged our team to design and build a centralized analytical platform capable of integrating data from all business areas, automating key processes, and providing leadership with a trustworthy, unified view of the company's performance.
.png)
Our team is building a comprehensive analytical solution on Microsoft Fabric, designed to serve as the single analytical backbone for all of Las Taperitas' business areas — from accounting and purchasing to agriculture, livestock, and beyond.
The foundation of the solution is a layered data architecture following the bronze–silver–gold model. Raw data lands in the bronze layer with minimal transformation, the silver layer applies cleaning, standardization, and business logic, and the gold layer exposes curated fact and dimension tables ready for reporting. This structure ensures traceability, repeatability, and a clear separation of concerns at every stage of the data pipeline.
Automated PySpark processes handle both historical loads and ongoing incremental updates, pulling data from transactional tables, APIs, and auxiliary sources. Given the complexity of Las Taperitas' operations, significant effort went into encoding business-specific logic: non-calendar fiscal years, monthly close periods, cost center hierarchies, and master key relationships all required careful modeling to reflect how the business actually runs.
One of the more nuanced challenges was monetary conversion. The solution applies specific rules to convert amounts into ARS, USD, EUR, and MEP dollar values depending on the product, period, and exchange rate type — a requirement that reflects the realities of operating in Argentina's financial environment.
Data quality controls were built into the pipeline to detect inconsistencies between facts and dimensions — particularly around cost centers, products, and master keys — before they reach the reporting layer. Auxiliary tables were also developed to support processes such as fixed asset depreciation, accumulated amortization, and fiscal period calculations.
The result is a scalable analytical foundation that reduces manual workload, strengthens internal controls, and gives management reliable, timely data to support decisions across every area of the business.
Data flows through three structured layers: raw ingestion, cleaned and standardized data with business logic applied, and curated analytical tables ready for consumption. This layered approach ensures full traceability and simplifies future maintenance and expansion.
Incremental and historical data loads run automatically, sourcing data from transactional databases, APIs, and auxiliary files — eliminating the manual extraction processes that previously burdened the team.
Fact and dimension tables cover the full breadth of Las Taperitas' operations: purchases, sales, accounting movements, products, suppliers, cost centers, fiscal periods, and exchange rates.
The model accommodates non-calendar fiscal years, monthly close periods, and month-end update processes — reflecting the specific accounting structure of the business rather than forcing a generic calendar model.
Automated rules handle monetary conversions to ARS, USD, EUR, and MEP dollar, applying the appropriate exchange rate based on product type, period, and conversion logic defined by the business.
Systematic validations detect mismatches between facts and dimensions — such as unmatched cost centers, missing products, or broken master keys — before inconsistencies reach the reporting layer.
Dedicated tables support complex calculations including fixed asset depreciation schedules, accumulated amortization, and fiscal year boundaries, enabling accurate financial reporting across reporting periods.
When source data changes retroactively, controlled reprocessing removes and reloads only the affected periods, preserving historical integrity while keeping the data model up to date.
Curated gold-layer tables feed Power BI reports for each business area, giving management a reliable view of balances, volumes, purchases, sales, and results — broken down by period, area, and any other dimension relevant to the business.
The architecture is designed to grow with the business — new data sources, business rules, areas, and reports can be incorporated without restructuring what already exists.
