This is a public Topic for those getting started with the Community and Quantexa Platform - content posted here will be visible to all.

Entity Resolution tooling and best practices

Rob_French Posts: 52 QUANTEXA TEAM
edited March 7 in Getting Started

Entity Resolution (ER) has always been the heart of any Quantexa solution and plays a significant role in the successful outcomes delivered.  

As we continue to improve this capability, we are working to make Entity Resolution clearer and lower the barrier to entry for all users.  

We are happy to announce we’re introducing new and improved documentation, focusing on how to understand your Entity quality and leverage new ER tooling through best practices and guides. 

Why Entity quality? 

Understanding and assessing the quality of Entities within any solution is often complex as it requires specialized skills, knowledge and resources such as: 

  • the ability to write Scala queries, 
  • knowledge of the batch build output,
  • time as investigations and queries often have long elapsed times, 
  • expertise in understanding and interpreting the results as well as recommend resolutions. 

This means that quality assessment historically could only be carried out by those with the required technical expertise, which could lead to investigations taking longer than necessary. 

What’s new? 

An earlier release of the Statistical Profile Testing (SPT) framework and the Bad Entity Analyzer (BAE) reduced the requirement level for both writing Scala and knowledge of the batch output. Although having these tools made assessing Entity quality much easier, we discovered a way to streamline the process even more. In the Quantexa 2.4 release, we integrated the Entity Explorer and Entity Quality Score into the Entity Resolution capability, making the visibility and overall understanding of the Entity quality a focus.   

You said; we did 

Whilst the introduction of new tooling is key, based on your feedback we've also identified the need for better documentation on Entity Resolution best practices, specifically on how to get the most of the features associated with tuning and analyzing Entity Resolution. We are pleased to say we have created a whole new Entity Resolution tuning best practice space within Quantexa Documentation in the Quantexa 2.4 release section.  

The new section will help offer advice and guidance through the process to help understand: 

  • Why is it important to validate an ER run? 
  • What are the signs of poorly resolved Entities? 
  • How do I use Quantexa tooling to simplify ER tuning? 

The documentation contains dummy examples of how to use the Entity Quality Scoring, Entity Explorer and Entity Lab to find, understand and propose suggestions on how to fix some common issues discovered. Watch out for future documentation on additional tooling and guidance around Entity Resolution processes as the site updates alongside Quantexa releases. 

Additional Resources

  • For more information on Entity Resolution check out our blog on Using base Entity Resolution configuration for the first time.
  • Whilst Quantexa's standard Compounds will support most use cases, there are situations where specific Compounds will be needed. See our article What Makes a Good Compound giving guidance and considerations to help with Compound creation.
  • Entity Resolution (ER) and good Entity quality underpins all Quantexa deployments. Therefore, we recommend that ER is monitored and reviewed through all stages of development. Our Entity Resolution general playbook describes the best practice process for monitoring and reviewing ER within Quantexa.

User Profiles
Getting Started Topic Owners
Ask our Topic Owners about anything you need to know to kick-start your Quantexa experience
VP Customer Success
Head of Customer Success Management - EMEA