User Guide

Welcome to the Trazadera Golden User Guide. This guide helps you understand how to use Golden for your day-to-day data quality and master data management tasks.

Who This Guide Is For

This guide is designed for:

Role	Primary Tasks
Data Stewards	Review duplicates, merge records, maintain data quality
Business Analysts	Query data, generate reports, monitor metrics
Data Managers	Configure entities, set up workflows, manage users
Operations Teams	Monitor tasks, troubleshoot issues, track performance

Quick Navigation

I want to...	Go to...
Understand the deduplication process	Understanding Deduplication
Review and merge duplicate records	Working with Buckets
Search for specific records	Finding Records
Monitor data quality	Monitoring and Reports
See a complete workflow example	Examples

Understanding Deduplication

Golden uses a multi-step process to identify and resolve duplicate records:

The Deduplication Pipeline

CODE

Source Data → Load → Index → Classify → Review → Merge → Golden Record

Step	What Happens	Your Role
Load	Data is imported from source systems	Configure sources
Index	Similar records are grouped into "buckets"	Define matching rules
Classify	System determines if records are duplicates	Set classification thresholds
Review	Uncertain matches are flagged for review	Review and decide
Merge	Confirmed duplicates are merged	Approve or customize
Golden Record	Best data is selected as the master	Define selection rules

Key Concepts

Concept	Description
Entity	A deduplication project (e.g., "Customers", "Products")
Bucket	A group of potentially duplicate records
Golden Record	The authoritative master record created from merged data
Classification	System's confidence level (DUPLICATES, REVIEW, UNIQUE)

Working with Buckets

Buckets are the heart of the deduplication workflow. Each bucket contains records that the system believes might be duplicates.

Bucket Classifications

Classification	Meaning	Action Required
DUPLICATES	High confidence these are duplicates	Auto-merged (or review if preferred)
REVIEW	Medium confidence - needs human review	Manual review required
UNIQUE	System believes this is not a duplicate	No action needed

Reviewing Buckets

When reviewing a bucket, you'll see:

All records in the bucket side-by-side
Match scores showing why records were grouped
Field comparisons highlighting differences
Golden record preview showing merged result

Making Decisions

Decision	When to Use	Result
Merge	Records are confirmed duplicates	Creates single golden record
Disconnect	Records are NOT duplicates	Separates into individual buckets
Skip	Need more information	Leaves for later review

Best Practices for Review

Start with highest-confidence REVIEW buckets
Use consistent decision criteria
Document unusual decisions with comments
Take breaks during long review sessions to maintain accuracy

Finding Records

Golden provides multiple ways to search for records.

Search Methods

Method	Best For	Example
Quick Search	Known identifier (email, ID)	`[email protected]`
Advanced Search	Multiple criteria	Name + City + Date range
Bucket Search	Finding specific bucket	Bucket ID `B-12345`
Filter by Status	Workflow management	All REVIEW buckets

Search Tips

Exact match: Use quotes for exact phrases: "John Smith"
Partial match: Use wildcards: john* or *@example.com
Multiple fields: Combine criteria: email:john* AND city:Boston

Working with Golden Records

Golden Records are the authoritative, deduplicated master records.

What Makes a Golden Record?

The Golden Record is created by:

Selecting the best value for each field from source records
Applying business rules (e.g., prefer most recent, prefer specific source)
Tracking lineage back to original source records

Golden Record Lifecycle

Status	Description
Active	Current master record
Updated	Modified by new source data or manual edit
Merged	Combined with another golden record
Archived	Soft-deleted, can be restored

Viewing Golden Record Details

Each Golden Record shows:

Master data: The consolidated field values
Source records: All contributing records with lineage
History: Changes over time
Quality scores: Data completeness and confidence

Monitoring and Reports

Key Metrics Dashboard

Metric	What It Shows	Target
Total Records	Records loaded from sources	Varies
Unique Records	Confirmed non-duplicates	Higher is better
Duplicate Rate	Percentage of duplicates found	Depends on data quality
Review Queue	Buckets awaiting manual review	Lower is better
Merge Rate	Daily/weekly merges completed	Track trends

Common Reports

Report	Purpose	Frequency
Duplicate Summary	Overview of deduplication results	Daily/Weekly
Source Quality	Data quality by source system	Weekly
Steward Activity	Review decisions by user	Weekly
Trend Analysis	Duplicate rates over time	Monthly

Setting Up Alerts

Configure alerts for:

Review queue exceeding threshold
Unusual spike in duplicates
Task failures
Data quality drops below threshold

Common Use Cases

Building a Customer MDM System

Goal: Create a single view of customer across all systems

Steps:

Create customer dataset with required fields (name, email, phone, address)
Set up sources (CRM, e-commerce, support systems)
Configure indexer for email, phone, name matching
Define classifier with weighted comparisons
Create merger strategy (prefer most recent, most complete)
Configure entity with AUTO_DUPLICATES mode for steward review
Schedule automatic synchronization
Monitor duplicate resolution progress

Relevant Guides: Entities, Resources, Golden Records

Implementing Data Integration Pipeline

Goal: Load and clean data from external sources

Steps:

Define source dataset matching source schema
Create source resource (JDBC, file, or API)
Define target dataset for Golden Core
Create transformation mapping fields
Create pipeline for data cleaning
Configure table for storage
Execute load operation
Monitor task completion

Relevant Guides: Resources, Tables, Tasks

Setting Up User Access Control

Goal: Ensure appropriate access to data and functions

Steps:

Configure authentication method (SSO or internal)
Define custom roles for your organization
Assign appropriate permissions to roles
Create user accounts with assigned roles
Generate access tokens for integrations
Configure entitlement filters for data privacy
Monitor and audit access regularly

Relevant Guides: Security

Manual Duplicate Review Process

Goal: Review and resolve uncertain duplicate matches

Steps:

Synchronize entity to create/update buckets
Query REVIEW-classified buckets
Examine records in each bucket
Make merge or disconnect decisions
Track progress via statistics
Generate reports on data quality improvements

See detailed walkthrough: Example: Manual Deduplication Workflow

Examples

Detailed step-by-step examples to help you get started:

Example	Description	Skill Level
Customer Data Integration	Complete ETL pipeline from CRM to Golden	Intermediate
Manual Deduplication Workflow	Step-by-step guide for reviewing duplicates	Beginner