All projectsCase study

Commercial Target Management

Multi-source business application: targets, prior-year comparison, planning, salespeople, summaries, exports and secured synchronization.

PythonFlaskMariaDBSQL ServerOracleHFSQL
70245/24515 min
01

Context

A commercial management application had to consolidate several store families and source systems into one interface without exposing the ERPs directly to the web layer.

02

Problem

Commercial sources are heterogeneous, publication schedules differ and some business rules require distinguishing missing data from a genuine zero. The interface must also remain usable when a remote source is temporarily unavailable.

03

Solution

Flask/Jinja + MariaDB architecture based on local caches and ERP connectors decoupled from the web layer. Scope includes store targets, prior-year comparison, planning, salespeople, multi-store summaries, business indicators, exports and controlled synchronization.

04

Key lesson

A reliable multi-source pipeline must be idempotent, observable and tested on every run.

MISSION Commercial Target Management
MY ROLE Development & data integration
DURATION Projet évolutif
ENVIRONMENT Python · Flask · MariaDB
STATUS Production + iterations
01 · BEFORE

Starting point

Multiple commercial sources, prior-year rules and KPIs spread across several systems.

→
02 · AFTER

What changes

One business interface powered by local caches and separated synchronization jobs.

PROJECT FLOW LIVE
01 ERP / DWH Sources
02 Connecteurs Lecture seule
03 MariaDB Caches
04 Python Métier
05 Dashboard Restitution
Under the hood

How the project actually works

Architecture, constraints, deliverables and real impact.

04

Main challenge

Reconcile several source systems, publication schedules and business rules while keeping the process auditable.

01 Remote ERP / DWH
02 Read-only connectors
03 MariaDB caches
04 Python business rules
05 Flask / Jinja / exports
05

Deliverables

  • Store and salesperson targets
  • Prior-year comparison
  • Salesperson planning
  • Multi-store summaries
  • Excel / CSV exports
  • Controls and audit trail
06

Impact

The web layer stays decoupled from the ERPs. Imports can be checked or replayed while the interface consumes validated local data.

Designed for real-world use
07

What this project demonstrates

I start from the problem and business flow, then aim for an observable, testable and maintainable solution rather than a demo effect.