Bot Food Lab

A small experiment in structured public information.

This site explores what kinds of factual web information are repeatedly useful to automated systems and to people.

Experiment status: Phase 5 — factual, semantic, and dynamic commercial-information datasets.

Experiments

Experiment A — Popularity / public attention

Time-stamped public popularity data that changes over time.

SixTONES YouTube — latest 30 videos and view counts

Experiment B — Local rules

Municipal rules, disposal methods, eligibility conditions, and effective dates.

How to dispose of lithium-ion batteries in Hamamatsu City, Japan

Experiment C — Compatibility / successor mapping

Discontinued products, official successors, specifications, and change history.

Panasonic FL4ENW/1P → FL4ENWF31P

Experiment D — Content taxonomy / semantic analysis

An anonymized semantic classification of 30 Japanese web essays.

Anonymous essay semantic taxonomy

Experiment E — Dynamic commercial pricing

A timestamped current-price snapshot for five fixed business hotels under one fixed search condition.

Hamamatsu business-hotel price snapshot

Method

Each test page favors primary sources, explicit verification dates, structured fields, and machine-readable data. Different information types are added as separate experiment groups.

Principles

Information should be factual, attributable, easy to parse, and updated when necessary. The site avoids private information and does not republish third-party text as a substitute for verification.

Structured schedule examples