Iranian Contact Data Normalization Pipeline

Problem

Raw phone exports contained inconsistent formats, invalid and landline values, multiple numbers per person, and duplicates that reduced downstream communication quality.

Users

People working with phone-contact exports and messaging workflows.

My Role

Independent data-product designer and developer; end-to-end ownership.

Key Decisions

Normalize and validate before selecting the usable contact.
Filter invalid/landline records explicitly instead of silently mixing them into output.
Prioritize operators consistently when multiple mobile numbers exist.

Delivery & Technical Approach

Built the Python package/CLI, schema validation, Iranian phone validation, pandas processing, normalization, filtering, deduplication, operator classification, and export.
Tested the operational workflow and generated Excel/messaging-ready output.

PythonpandasETLRegex validationCLICSV / Excel

Product / System Flow

01
Raw exports
02
Normalization
03
Phone validation
04
Filtering & deduplication
05
Operator-aware selection
06
CSV / Excel output

Outcome / Impact

Produced clean output ready for Excel or messaging workflows.

Constraints & Boundary

Archived repository snapshots do not necessarily expose every behavior of the operational version.
Data-quality claims stay tied to the tested workflow and documented validation rules.

Scope note: Describe processing behavior through the documented validation rules; do not imply guaranteed throughput.

Reflection & Next

Data quality works best as an explicit product contract with repeatable rules.
Repeatable validation makes downstream outputs easier to trust.

Next: Add regression fixtures for Iranian-number edge cases and keep validation rules explicit as the workflow evolves.