README.md

metadata

dataset_info:
  features:
    - name: dataset_id
      dtype: string
    - name: dataset_name
      dtype: string
    - name: source_url
      dtype: string
    - name: supported_entities
      sequence: string
    - name: supported_events
      sequence: string
    - name: description
      dtype: string
    - name: last_updated
      dtype: timestamp[ms]
    - name: coverage_timeframe
      dtype: string
    - name: created_at
      dtype: timestamp[ms]
  splits:
    - name: train
      num_bytes: 3419
      num_examples: 12
  download_size: 8586
  dataset_size: 3419
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-4.0

Crisis Prediction Datasets

Dataset Card for dwb2023/crisis_prediction_datasets

Purpose

A catalog of datasets supporting crisis prediction and response analysis, with particular focus on transportation and supply chain disruptions. This catalog tracks datasets that can be used to model temporal dependencies, test counterfactual scenarios, and analyze cascading failures in crisis situations.

Intended Use

• Primary Use: Research and evaluation of crisis response datasets
• Intended Users: Data scientists, data engineers, and ML engineers working on crisis prediction systems
• Out of Scope: Real-time data ingestion or operational deployment

Schema Description

The catalog uses the following schema to track and document datasets:

Field Name	Data Type	Description
dataset_id	string	Unique identifier for the dataset in the catalog (e.g., "OPENSKY_2020", "GDELT_2020_Q1")
dataset_name	string	Human-readable name of the dataset (e.g., "OpenSky Network", "GDELT Event Database")
source_url	string	URL or reference point for accessing the dataset
supported_entities	sequence[string]	List of entity types contained in the dataset (e.g., ["Airport", "Flight", "Policy"])
supported_events	sequence[string]	List of event types the dataset can track (e.g., ["FlightCancellation", "HubClosure"])
description	string	Detailed description of the dataset's contents, format, and relevance to crisis prediction
last_updated	timestamp[ms]	When the dataset itself was last updated, in milliseconds since epoch
coverage_timeframe	string	Time period covered by the dataset (e.g., "2020-03-20 to 2020-03-27")
created_at	timestamp[ms]	When this catalog entry was created, in milliseconds since epoch

Example Entry

{
    "dataset_id": "OPENSKY_2020",
    "dataset_name": "OpenSky Network",
    "source_url": "https://opensky-network.org/api",
    "supported_entities": ["Airport", "Flight"],
    "supported_events": ["FlightCancellation", "FlightDelay"],
    "description": "Real-time & historical ADS-B flight tracking data",
    "last_updated": "2025-01-30",
    "coverage_timeframe": "2020-03-20 to 2020-03-27",
    "created_at": "2025-01-31"
}

Key Features

• Tracks datasets supporting temporal graph analysis
• Maps datasets to specific entity and event types
• Documents coverage periods aligned with crisis events
• Enables dataset comparison and selection for analysis

Usage Caveats

• Catalog is for research planning purposes only
• Does not include actual data storage or automated ingestion
• Focus is on iterative research evaluation of datasets

Implementation Details

Query Examples

-- Find datasets containing flight data
SELECT dataset_name, source_url 
FROM dataset_catalog 
WHERE supported_entities LIKE '%"Flight"%';

-- Get datasets covering March 2020
SELECT dataset_name, description 
FROM dataset_catalog 
WHERE coverage_timeframe LIKE '%2020-03%';

-- List recently updated datasets
SELECT dataset_name, last_updated 
FROM dataset_catalog 
ORDER BY last_updated DESC;

Key SQL Operations

-- Create new dataset entry
INSERT INTO dataset_catalog (
    dataset_id, dataset_name, source_url,
    supported_entities, supported_events,
    description, coverage_timeframe
) VALUES (
    'GDELT_2020_Q1',
    'GDELT Event Database',
    'https://www.gdeltproject.org/',
    '["Airport", "Policy", "Organization"]',
    '["PolicyChange", "BorderClosure"]',
    'Global event database tracking media-reported events',
    '2020-01-01 to 2020-03-31'
);

-- Update dataset coverage
UPDATE dataset_catalog
SET coverage_timeframe = '2020-01-01 to 2020-06-30'
WHERE dataset_id = 'GDELT_2020_Q1';

Citation

This catalog was developed as part of research into crisis prediction and response systems, focusing on the March 2020 COVID-19 disruptions to global air cargo operations.