liveBook · Manning

Language model

As explained in Everyday Data Visualization

A large language model is a machine-learning model trained on extensive text data to process and generate human language.

A language model is a machine-learning model trained on extensive text data to process and generate human language. Here, it becomes a practical source of realistic test data when production data does not yet exist. section 9.1.2

Shape generated test data so prototypes resemble the eventual product:

  • Specify realistic columns, including dates, times, names, or addresses. section 9.1.2
  • Describe realistic values and constraints, such as geography, future dates, or valid rating ranges. section 9.1.2
  • Request realistic distributions, such as subscription spending or house prices, and refine the prompt when needed. section 9.1.2

Survey prototype data

ChatGPT generated two respondents answering three survey questions, returning respondent identifiers, question text, and numerical responses in a comma-separated structure. section 9.1.2

How Everyday Data Visualization covers it

How deep it goes
Defines the model through a practical use case but stops short of explaining training, architecture, or language-generation internals.
Examples
Uses a ChatGPT-generated survey containing two respondents, three questions, and numerical rating responses.
What it assumes
Assumes a desired data structure, column names, questions, and constraints that can be communicated in a prompt.
What you can do after
Apply via ChatGPT to generate structured, realistic test data for prototypes before production data becomes available.
How it connects
Connects language models with generative AI, ChatGPT, test data, surveys, and data visualization prototypes.

Also explained in

Compare how 67 books explain language model →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Everyday Data Visualization. The section links point to the parts of the book it draws on.