Topical analysis – also known as thematic analysis, topic modeling or qualitative data coding – is at the very heart of our text analytics platform. Put simply, it is a classification under which words are grouped, based on a shared concept. In Relative Insight, we use it as a powerful tool to help derive insights from qualitative data, by comparing text and surfacing differences that provide true business intelligence.
The fluidity of language
Language is not a constant – it develops and changes based on how we use it. In the digital age, we often see new words being created, depending on what is said and how it is understood.
The new topic sets are a lot more tangible and commercially relevant, to reflect the nature of and the realities of modern communications and consumer behavior. This ensures that you get the most out of your text analysis in Relative Insight Explore.
Topics such as “Fitness”, “Public safety”, “Investment”, “Streaming” and “Mental Health” as well as “Cybersecurity” will be included when you analyze new projects. We have also introduced another emotion – “Surprise” – to support you in gaining a deeper understanding of the sentiment within your text data.
Putting the right label on things
Our data scientists began to overhaul Relative Insight’s set of semantic topics earlier this year. Relative Insight’s roots are in academic research, so some of the previous phrases for labeling themes in the topic analysis seemed opaque and did not fully encompass the realities of modern language use.
Since March 2022, over 250 topics have been renamed to improve the modeling capabilities and business applications in Relative Insight, while removing obstacles and hindrances to increasing clarity.
For example, “Relationship: Intimacy and sex” became simply “Love” and “Discourse bin” has been renamed to “Slang.” Of course, you can still rename the topic yourself, based on what fits the requirements of your analysis.
These updates will provide your insight teams with a refined topical analysis view of their text-based data, offering a better understanding of what their customers, employees or competitors are saying, in a more efficient manner.