Commercial & inventory
A single view of sales and margin
Sales, purchases, margin, stock position and target attainment consolidated in a single source.
All the stores’ numbers in one place, and an app to act on them.
I designed and built a complete data platform for Grupo Althoff — a supermarket chain in southern Santa Catarina. It brings together the stores’ and the distribution center’s systems, consolidates over 1 billion records in one place and delivers decisions two ways: dashboards where each manager sees their own store, and a proprietary app used every day by the stores.
Role · Architecture, data engineering, BI, ML and full-stack development
Scope · End-to-end project · in production
In short, for decision-makers
The problem
Althoff’s information was locked in systems that didn’t talk to each other — ERP, HR, DC inventory, spreadsheets — so building any report meant people exporting and cross-referencing data by hand, always late.
What LNData did
We brought all those systems into a single base and delivered decisions two ways: dashboards where each manager sees their own store, and a proprietary app where stores act on the data day to day.
The result
The chain now decides with fresh numbers — previous-day sales, inventory, losses, targets — and stores handle receiving, control expiry and place orders in the same system, with no loose spreadsheets.
10
stores + DC + administration
~1,000
employees
~17k
products in the mix
since 2022
analytical history
+1B
records consolidated in the data warehouse
77
automated data pipelines
16
governed BI dashboards
2
delivery channels (BI and operational app)
Architecture
The platform integrates multiple source systems, versions and conforms the data into an analytical data warehouse and delivers the result through two channels: BI dashboards and a proprietary operational application.
Source systems
- Commercial ERP
- HR system
- Distribution center WMS
- Occupational clinic
- APIs and network spreadsheets
- Competitor prices
Data platform
Pipeline orchestration
Analytical data warehouse
Parquet datalake · ML and NLP · containerized
Consumption
- BI dashboards · per-store access
- Proprietary operational application
- Email reports · XLSX / PDF
Areas covered
From commercial to the distribution center, every business area integrated into the same data platform.
CRM & loyalty
Purchase behavior
Customer analysis and RFM segmentation to guide campaigns and relationships.
Losses & breakage
Loss control per store
Losses and breakage by store and department, with reports sent automatically to managers.
HR / People Analytics
People and productivity
Headcount, turnover and productivity cross-referencing time clock with operational results.
Distribution center
DC productivity
Picking, receiving and replenishment closely tracked, with indicators updated throughout the day.
Finance & accounting
Foundation for financial management
Accounts payable and receivable and accounting entries organized as an analytical base.
Applied intelligence
Machine learning, AI and automated collection in production — not as a proof of concept, but generating decisions day to day.
Demand forecasting (ML)
Time-series models by store and product generate an order suggestion ready for promotional campaigns.
Product matching (AI / NLP)
Semantic comparison of the in-house mix with competitors’, even when there is no common barcode.
Price intelligence
Automated collection of millions of competitor price observations to support commercial decisions.
Product · Althoff Toolbox
The application where stores act on the data
If BI is the analytical side (read-only), the Toolbox is the operational side: a proprietary web application where staff and managers act on the data — from receiving checks and expiry control to order suggestions and HR communications.
Expiry control
Expiry audits in a spreadsheet-like grid, with markdown price suggestions and a full trail.
Receiving verification
Verification of inter-store transfer invoices, with an auditable history of each adjustment.
Order suggestion
The sales forecast per campaign becomes a purchase suggestion ready to review and export.
Governance and access
Access by store and by role, with an audit record of every action taken.
FlaskVue 3VuetifyPostgreSQLPlotlyPWA
Impact
Data once trapped in silos — ERP, HR, WMS and network spreadsheets — now lives in a single analytical source of truth.
Daily decisions based on fresh data: sales up to the previous day and operational indicators refreshed throughout the day.
Each manager sees only their own store’s data, with end-to-end access governance.
Stores don’t just view: they act on the data — expiry, receiving and orders — through a proprietary application.
Platform stack
AirflowClickHouseSupersetPythonMachine LearningFlaskVue 3Docker
A similar data challenge?
From integrating siloed systems to putting ML and governed BI in production — tell us what you need to solve and we’ll come back with scope, timeline and deliverables.