This map shows geo-located tweets covering a 24 hour period, obtained using Twitter’s Streaming API. Each blue dot represents one tweet. The red line represents the KL-Selangor border. Click here to view the full-size image on Flickr.
Previous maps that we released were based on cumulative data from 2010, from tweets on socio-political topics. This map is based on all tweets regardless of topic. We were planning to collect a months’ worth of tweets before sharing the results, but the significant difference in just 24 hours makes it worth sharing now.
Tweets : 145145
Users : 24894
This can be broken down into:
Tweets : 103303
Users : 17414
Tweets : 24974
Users : 4800
Tweets : 16868
Users : 2680
Tweets : 61680 (59.70% of Malaysia)
Users: 10902 (62.60% of Malaysia)
Rest of Malaysia
Tweets : 41623 (40.29% of Malaysia)
Users : 6512 (37.40% of Malaysia)
Past experience has shown us that only 5-10% of tweets on a given topic are geo-tagged. However it is difficult to estimate the number of users based on tweets. 24 hours worth of data is also too small a sample. Still its worth calculating daily estimates at this point to see what sample size is best.
Based on the Malaysia Total stats, and assuming an average of 6 tweets/user we can calculate a simple estimate of the size of the population:
Min = (103303 / 10 * 100)/6 = 172172
Max = (103303 / 5 * 100)/6 = 344343
Therefore the estimated size of Malaysia’s active Twitter population is 172,172 – 344,343 users. Our last census of politicians’ followers in August found 594,893 active users, but that includes foreigners and Clones. We will continue to collect data and recalculate this estimate in 1-2 weeks time.
Update #1 (September 14)
Found a large number of Borneo tweets were included in the sample obtained from Twitter. This is due to an issue with their Streaming API. All calculations have been revised to reflect the change.