Steve Rubel points to a search engine for Twitter. I've been curious about what people talk about on Twitter (of course, I'd like to see something which does trend graphs over this data stream...). The best I could think about just now was to perform searches and look at the time range for what I assume is the most recent N posts. Below, I show searches and their start and end times - that is to say, the stamp that Twitter gives for the first and last result. Unless otherwise noted, I looked at 100 results. When there were fewer, the maximum number is shown. Note also that the indexing in this engine is not standard. For example, 'rain' matches 'train', 'brain', etc.
| Search | start - end |
|---|---|
| breakfast | 3 mins - 21 hours |
| sleep | 2 mins - 6 hours |
| tv | 3 mins - 14 hours |
| lunch | 3 mins - 20 hours |
| dinner | 13 mins - 13 hours |
| rain | 2 mins - 16 hours |
| coffee | 3 mins - 15 hours |
| iraq | 4 mins - 1 day |
| bush | 26 mins - [2 days, 80] |
| iran | 1 hour - [2 days, 30] |
| clinton | 11 hours - [2 days, 13] |
| obama | 2 hours - [1 day, 16] |
From this totally ad hoc and uncontrolled experiment - and this is not a criticism of Twitter - it appears that relatively speaking, Twitter is more a channel for annotating one's daily activities than it is to express opinions (at least, if we take politics as an example of a space in which people could express opinion). That is going to change. There are a number of institutions which are starting to use Twitter to channel hooks to their content - e.g. news sites either by themselves or via some third party mashing are starting to publish content to Twitter. This will likely impact the topical space of Twitter content.
Comments