Skip to main content

Table 3 Distribution of sampled Tweets, geo-filtered Tweets and Tweets filtered into one our Topics by MSA and overall for our experiment. These counts represent samples gathered from all methods and sizes included in our experiment over the 38 days in our field period

From: Sweet tweets! Evaluating a new approach for probability-based sampling of Twitter

Metropolitan statistical area (MSA)

Total tweets sampled

Total tweets geo-filtered into principal cities

Total tweets filtered in to one of our 8 topics

Tweets from sample missing user geography metadata

Chicago–Naperville–Elgin, Illinois–Indiana–Wisconsin

22,442,771

18,525,841

531,402

178

Atlanta–Alpharetta–Sandy Springs, Georgia

22,413,310

17,624,013

329,958

90

Phoenix–Mesa–Chandler, Arizona

22,607,714

17,191,865

640,358

60

Baltimore–Columbia–Towson, Maryland

22,435,821

16,087,382

446,228

29

Pittsburgh, Pennsylvania

22,424,971

11,099,119

393,554

309

Grand Total from All MSA Regions in the Experiment

112,324,587

80,528,220

2,341,500

666