UC Irvine
ML Repository
Theme

Twitter Geospatial Data

Download(179.6 MB)
Thumbnail

About

Seven days of geo-tagged Tweet data from the United States with exact GPS location and timestamp. Note that this is the full week of data that was sampled from Twitter. The 10,005,301 count mentioned in the introductory paper below refers to the weekday portion of the data (i.e., Monday through Friday). If you remove Saturday (Jan 12, 2013) and Sunday (Jan 13, 2013), then you will get the Monday through Friday portion that was analyzed in the paper. Variables Info: This dataset contains geospatial and timestamp data for one week worth of Tweets in the contiguous United States. The Tweets were created between January 12, 2013 and January 18, 2013. The exact location (i.e., longitude and latitude) and timestamp (hour, minute, second) of each Tweet was recorded. All timestamps are reported in central standard time in the format "YYYY-MM-DD HH:MM:SS". The geo-tag information was used to assign each Tweet to one of the four standard time zones (for details see Helwig et al., 2015). The data were collected by the CyberGIS Center for Advanced Digital and Spatial Studies at the University of Illinois at Urbana-Champaign. Details on the data preprocessing and analysis can be found in Helwig et al. (2015). Class labels: 1. longitude: exact longitude coordinate of Tweet (real valued) 2. latitude: exact latitude coordinate of Tweet (real valued) 3. timestamp: 20130112000000 = 2013-01-12 00:00:00 CST (integer) 4. timezone: 1 = Eastern, 2 = Central, 3 = Mountain, 4 = Pacific (integer)
Subject Area
Social Science
Instances
14,262,517
Features
4
Data Types
Multivariate, Spatiotemporal
Tasks
Classification, Regression, Clustering
Feature Types
Continuous, Categorical, Integer

Features

Introductory Paper

Analyzing spatiotemporal trends in social media data via smoothing spline analysis of variance
Nathaniel E. Helwig, Yizhao Gao, Shaowen Wang, Ping Ma. 2015.
Spatial Statistics

Additional Metadata

Keywords
Authors
Nathaniel E. Helwig
Yizhao Gao
Shaowen Wang
Ping Ma
Year Created
2015
License
CC BY 4.0