Data Quality Correction — Quebec Real Estate Contact

I need an experienced data quality specialist to correct and fully re-verify a real estate contact dataset covering the Quebec City market (Canada).

── DATASET STRUCTURE ──
Please read this before applying — it affects how you estimate the work.

The dataset has 1,500 rows in Google Sheets. It does not represent 1,500 unique contacts. Each row = one property lot. One owner can hold several lots and appear on multiple rows. After deduplication: ~760 unique property owners.

This means:
• Phase 2 verification covers ~760 unique owners, not 1,500 rows
• Phase 1 corrections must be applied to all affected rows for each owner
• Part of the work involves identifying and mapping these duplicates

Please factor this into your price and timeline estimate.

── CONTEXT ──
The dataset was previously enriched by a freelancer (phone numbers, emails, LinkedIn profiles, decision-maker names added to each owner record). An independent quality audit identified 47 anomalies across 5 categories:

• Category A — Contact marked "Available" but no phone provided (11 records)
• Category B — Owner / decision-maker / contact mismatches (10 records)
• Category C — Duplicate contacts / over-solicitation risk (6 groups)
• Category D — Missing or incomplete data (10 records)
• Category E — Invalid formats, OCR errors (8 records)

Current confidence score: ~55/100. Target after correction: 85/100 or above.

── WHAT I WILL PROVIDE ──
✔ Google Sheets link (1,500 rows / ~760 unique owners)
✔ Full Quality Audit Report (Word, ~10 pages) with every anomaly, severity level and recommended fix — shared with shortlisted candidates only
✔ Clear written instructions before you start

── SCOPE OF WORK ──
Two mandatory phases:

PHASE 1 — Fix known anomalies (from audit report)
• Correct all HIGH severity issues (15 records): owner/decision-maker mismatches, OCR-broken emails, invalid country codes, merged phone+email fields
• Resolve MEDIUM severity issues (22 records): incomplete records, homonym risks, inconsistent data
• Deduplication mapping: add a Unique Owner ID column, group all rows by owner, flag owners appearing on 5+ rows

PHASE 2 — Full independent verification (~760 unique owners)
Verify every unique owner — not just flagged ones. For each record check:
• Decision-maker name matches the legal owner or registered company representative
• Phone number is valid with correct country code (+1 Canada)
• Email is syntactically valid and linked to the person or company
• LinkedIn profile corresponds to the listed decision-maker

Mandatory sources (minimum): LinkedIn + Quebec Business Registry (REQ: registreentreprises.gouv.qc.ca) + company website when available. No phone verification required.

── DELIVERABLES ──
1. Corrected Google Sheet with:
— All anomalies fixed or flagged with a status note
— Unique Owner ID column
— Verification Status column per owner (Verified / Corrected / Flagged / Unable to verify)
— Correction log tab: row ID | owner | original value | corrected value | source | status

2. One-page summary: new anomalies found in Phase 2, unverifiable records and reasons, overall quality assessment

── YOUR PROPOSAL MUST INCLUDE ──
1. Your price for the full project (Phase 1 + Phase 2)
2. Your estimated delivery time
3. Your Phase 2 verification methodology
4. One or two examples of similar data quality projects

Note: proposals that do not reference the 1,500 rows / 760 unique owners structure will not be considered. Shortlisted candidates will be asked to complete a small paid test (10–15 records) before the full project is awarded.