Tables are having a renaissance after years of neglect in online content. Tables are assuming a bigger role in online content, displacing narrative text in some cases.
Web authors found tables troublesome to create. But machines find them easy, which is accelerating the structuring of information.
I recently posted on LinkedIn about an article that noted the prevalence of tables in chatbot answers. The study found that approximately 30% of ChatGPT responses included tables, approximately two or three times more than in traditional search results. Based on comments posted on LinkedIn, many others also noted the prevalence of tables.
LLMs can convert text narratives into tables, providing a concise and clear summary of verbose prose.
Google has released Text2Table functionality in their NotebookLM product. NotebookLM is a retrieval-augmented generation (RAG) tool that can summarize and extract information from user files. The Data Tables feature allows users to organize dispersed information instantly. “Key facts are often scattered, making manual compilation tedious. Today, we’re making that simpler with Data Tables.”
While organizing information into rows and columns is useful, tables can do much more to show relationships within the information.
A Python-based tool called “Great Tables” showcases options for automatically generating tables. The goal of Great Tables is to provide “affordances for structuring information for better legibility and how the package can be used to adorn the table with other structural elements.” Its creators note that most tools generate primitive tables that lack the rich nuances of hand-designed tables from the past. Now, sophisticated tables can be generated on the fly.
Tables have a bright future. They are easier than ever to create, and allow users to see a vast array of information at once to make comparisons and understand relationships.
— Michael Andrews