Tableau | AI & ML | Data visualisation | Tool strategy S2 E3: Bit - NLP (Natural Language Processing).
NLP is powerful, but if your data fields aren't aliased and described properly, your natural language search is just guessing.
- NLP works by converting words into numerical values a machine learning model can process, then layering linguistic structure back on top
- Tableau's Ask Data (built on the acquired ClearGraph technology) scans metadata first, then matches typed questions to fields, aggregations and date ranges
- Field aliasing lets you map business synonyms (sales, revenue, transactions) to the same underlying field so natural language queries resolve correctly
- Syntax (grammar and structure) and semantics (derived meaning) are distinct challenges, and ambiguous phrasing or slang still trips up NLP models
- Natural language interfaces shift the real bottleneck onto data curation and governance rather than removing the need for it
Future-proof your career https://n1d.io
| It’s an odd number episode so we’re back to Bits, episodes focussed on a technology. Today we talk about Natural Language processing, a subset field of Machine learning.
Show notes:• NLP: Machine learning Guide: (http://ocdevel.com/mlg/18)
• Example of a Corpus: Quora: Where can I get a large corpus of news articles? (https://www.quora.com/Where-can-I-get-a-large-corpus-of-news-articles)
• MODERN Business Intelligence: NLP: It’s not a Feature, it’s the Future! (https://medium.com/@johnt_16099/modern-business-intelligence-nlp-its-not-a-feature-it-s-the-future-40a63d1aeb61)
• Tableau: Natural language humanizes your data (https://www.tableau.com/reports/business-intelligence-trends/natural-language)
Feedback welcome on Twitter to Ravi at @scribblr_42 or Tim at @tableautim - or e-mail us, at [email protected] (mailto:[email protected])
