Skip to content

Watts-Lab/commonsense-data

Repository files navigation

Common Sense Data Repository

This repository contains data organized into various folders, each representing a table in the database.

Folder Structure

  • statements: Active statements running on the platform.
  • statementproperties: Properties of the statements for the design point experiment.
  • answers: Answers to the common sense questions uniquely identified by the sessionId.
  • experiments: Experiments conducted on the platform. Each experiment is uniquely identified by the experimentId.
  • individuals: Answers to the CRT and RMET questions uniquely identified by the sessionId.

Usage

The data is stored in chunks of 100 megabytes each. To extract the data in python for example, you can use the following code:

import pandas as pd

statements_csv = [f for f in os.listdir('statements') if f.endswith(".csv")]
statements = pd.concat([pd.read_csv(f'statements/{f}') for f in statements_csv])

It is possible to join the data from different tables using the sessionId or experimentId columns.

License

To be determined.

About

Automated releases of public data from our common sense experiments

Resources

Stars

Watchers

Forks

Languages