Datasets:

ariels
/

pest_twitter_news_classification

This dataset classifies Twitter and News posts about two plant pests, Spotted Lanternfly and Tuta absoluta into 10 categories (multi) and 3 categories (locational) defined below:

Categories

All categories

Spread/extent: Known locations of a pest (e.g. list of districts already known to have the pest, already quarantined locations), and/or broad descriptions of new locations, secondhand descriptions of pest observations
Management/control: Control measures that readers can take or that are being taken by authorities or farmers
Damage/costs: Agricultural, ecological and/or resulting economic losses (from both lost crops and lost trade)
Direct sighting: A new direct sighting of the pest (ie. equivalent of an iNaturalist observation); direct discussion of symptoms on a plant (or action taken (squash, stomp...); Conversations by individuals about having seen the pest or someone responding to confirm that the symtoms or images indicates that it was a sighting.
Government action/biosecurity: Decisions, mandates, bans/embargos taken directly by governing bodies (countries, lower admin units (e.g. state/county), or multi-national groups e.g. UN)
Public awareness: Informative about pest (Tuta absoluta is a....), lists of factors/descriptors (e.g., 5 things to know about SLF...), general warnings about the risk of the pest
Research: Proposed and ongoing research, scientific publications, and/or research findings
Funding: Money being put towards research, management, and/or control of the pest
Reactions to control: Reactions to management/control: pushback, appreciation, collateral damage
Irrelevant: Content that does not fit into the above categories, off-topic

Locational classification

Spread/extent: Known locations of a pest (e.g. list of districts already known to have the pest, already quarantined locations), and/or broad descriptions of new locations, secondhand descriptions of pest observations
Direct sighting: A new direct sighting of the pest (ie. equivalent of an iNaturalist observation); direct discussion of symptoms on a plant (or action taken (squash, stomp...); Conversations by individuals about having seen the pest or someone responding to confirm that the symtoms or images indicates that it was a sighting.
Other: Content that does not fit into the above categories

This dataset was produced as part of publication: Tateosian, Laura G., Ariel Saffer, Chelsey Walden-Schreiner, and Makiko Shukunobe. “Plant Pest Invasions, as Seen through News and Social Media.” Computers, Environment and Urban Systems 100 (March 1, 2023): 101922. https://doi.org/10.1016/j.compenvurbsys.2022.101922.