Refinement, More Customization, and New Models – BlueML V1.4.0 Arrives

Written by Lorcan Archer, Explorance.

A screenshot of the latest version of BlueML, Explorance's HR-trained comment analysis solution

July 2021 saw the landmark release of BlueML, Explorance’s unique machine learning-powered comment analysis solution.

BlueML is unique, being HR-trained to turn employee feedback comments into decision-grade intelligence. Capable of analyzing open comments at scale, BlueML ensures that each employee’s voice is heard, helping to enable meaningful feedback, engagement, and retention while enhancing speed to impact.

After an excellent reception to BlueML at Explorance’s Bluenotes GLOBAL and MTMImpact Symposium conferences this summer, Explorance has now delivered the first major updated to BlueML, refining and expanding the impressive capabilities of BlueML.

Commenting on the update to BlueML version 1.4.0., John DeVille, Principal Consultant at Explorance, noted the increased focus on customization in this update, and how this was delivered in response to client needs.

“In addition to more intuitive and powerful analytics, much of the new functionality in BlueML 1.4 was in response to our clients’ needs for customization. This is both in terms of how employees’ feedback is analyzed and how BlueML interprets and categorizes that organization’s terms and acronyms during the analysis.”

“With this update, customized categories and terms based on the unique needs of your company allow you to maximize the amount of info you can extract from the pool of available data,” John noted. “This increases the validity of your analysis and the credibility of your findings with your stakeholders.”

BlueML 1.4.0 – Now Featuring Upgraded UX and UI

A major general enhancement to BlueML present in BlueML 1.4.0 is an across-the-board enhancement to UI and UX. This will be instantly visible to users once they log into the updated version. The general visual presentation and graphs have been optimized to better facilitate the data exploration process.

Crucially, the summary zone of the dashboard can now be easily personalized, permitting users to adjust, add, or remove the displayed widgets – depending on your own focus or those factors that the institution is particularly interested in investigating.  

What BlueML 1.4.0 delivers can broadly be categorized into three broad sections. These are:

  • Deepened Analysis Options
  • Categorization Tweaking and Input
  • New models

How does BlueML 1.4.0 Deliver Deeper Analysis?

More in-depth analysis is possible thanks to these new aspects of BlueML.

  • Categorization refinements for greater precision – The former 2-level categorization (category > attribute) has been further refined to include more topics and even sub-topics under certain categories. This deepened branching out of the categories, permitting an even more precise description of what is being analyzed.
  • Addition of widgets to summarize results for quick insights – Widgets encapsulate results to supply quick, actionable insights. The three widgets introduced in BlueML 1.4.0 highlight the most positive topics, the most negative topics, as well as the most discussed topics overall.
  • Ability to drill-down by topic or sub-topic – The categorization drill-down has been enhanced to allow you to target a specific category, topic within a category, or sub-topic. This new type of drill-down allows you to highlight comments that belong to any topic whether they are top level categories or lowest level topics as well as show positive and negative sentiments where applicable.
  • Unique commenter count – You may now specify a unique identifier per comment allowing you to display not only how many comments were analyzed overall but also how many unique commenters contributed to them.
  • Ability to submit feedback on the analysis – Users can be permitted to give their own feedback on the analysis results. This allows the BlueML team to work off this feedback and continuously improve the platform. Ultimately, this will improve BlueML’s overall value and capabilities.

How Has Categorization Improved with this Version of BlueML?

These new aspects to BlueML ensure that the solution is more personalized and allows users to tailor their use of it to their own specific needs. BlueML already understood the unique language of HR, you can easily teach BlueML 1.4.0 to understand the unique language of your organization.

  • Giving meaning to localized abbreviations and special keywords – Abbreviations can be tricky! We know that there many abbreviations used within your organization that are not a standard dictionary term – but have a special and important meaning to your organization or institution. Leveraging BlueML Dashboard’s new customization area will allow you identify these special terms used in your organization and apply any common dictionary synonym. BlueML will do the work of interpreting them and categorizing them accordingly – helping your in-house language become more closely reflected in how BlueML works.
  • Customizing categories for specialized analysis – Now you can tweak BlueML’s categorization by renaming the default topics and even creating your own custom topics and subtopics from any item within the BlueML model. With this powerful flexibility, users can mix and match topics to highlight brand new themes and summaries that are especially meaningful to your organization’s stakeholders.

What New Models Does BlueML 1.4.0 offer?

BlueML supports specific models that are designed to consume and analyze comments related to a specific aspect of the student and employee journeys. These models help to detect and categorize the emotions embedded in open-ended responses.

Note Beta models are available on demand, so contact your account manager for more details.

  • Student Learning Categorization (SLC) model - Beta - The SLC model categorizes the evaluation of a course by a student at a higher education institution into themes, categories, and attributes as well as by sentiment. Use this model to analyze comments from internal surveys as well as external sources.

  • Emotions model – Beta - The Emotions model is a supporting model that detects the emotion conveyed in a text, such as anger. Seven different emotions can be detected with this release. 

That does it for this release! If you’re new to BlueML, we invite you to follow the below link to learn how BlueML can deliver for your organization.


How BlueML provides indepth analysis of your open text comments in seconds

BlueXComment analysisExperience Management

Get in touch with us about this article.

Stay connected
with the latest products, services, and industry news.