Network Change Registry

Traceable, consistent network-change history, feeding the KPI programme that cut external complaints by 15%.

  • Oracle
  • SQL
  • Python
  • Power BI
2020–2021

Details abstracted to protect client confidentiality.


Context

In a mobile network, configuration changes land daily, from many hands. Without a single record of what changed, where and when, every performance investigation starts with archaeology, and undocumented changes surface later as customer complaints.

My role

As Analytics & Planning Engineer at Ooredoo, I designed the internal database that centralised network-change tracking, and built the network-performance KPIs and reporting on top of it.

Approach

  1. Mapped how changes were actually being recorded across teams, and where history was being lost.
  2. Designed a SQL schema that centralised change tracking: one registry, controlled fields, one consistent structure.
  3. Built network-performance KPIs on Oracle DB, read against the change history.
  4. Automated the dashboards and reporting with Python, Pandas and Power BI, so the numbers arrived without manual assembly.
FIG 01 · pattern-level view · one registry, many consumers · client specifics removed

Governance decisions

  • One registry, not many spreadsheets. A single source of truth for change history is the precondition for every other control.
  • Controlled fields. Consistent structure and constrained values cut errors and duplicates at the point of entry, the cheapest place to fix them.
  • Traceability as the product. Every change carries enough context (what, where, when) to be found again; troubleshooting becomes lookup rather than recollection.
  • Documentation inside the workflow. Recording a change is part of making it, not an afterthought to chase later.

Outcome

The KPI programme this fed cut external complaints by 15%. Change history became traceable and consistent, documentation improved, and errors and duplicates fell: the quiet, structural work the visible number depended on.