Wanted to start a thread here in case anybody wants to comment on the launch of SickCity. Comments, ideas, criticisms (constructive, of course), tangential musings? Post em here.
I think SickCity is pretty cool, but I just wanted to highlight another tool that mines Twitter for tweets pertaining to any topic and then applies a sentiment algorithm to it in order to understand how people feel about the product or topic. It was developed by Microsoft's PR agency Waggener Edstrom. Here's the link:
This could be a great tool. It does however need some correlation to reality. Data collected by the CDC is generally delayed. What Google did was correlate their trend data with the CDC's data. This showed that Google's was a leading indicator for flu trends.
The same thing could be done with Twitter data. I know that their are some surveillance systems in place in various cities that collect flu data in near real time from hospital emergency rooms. This would likely be a good place to look for correlations.
Public Health is not likely to accept data that does not have some proven statistical basis.
You're absolutely right bujanga. Would you be willing to help guide us in getting some sort of correlation like that in place? That's a crucial next step in the development of the tool.
Twendz--Twitter sentiment mining tool
By justindaveyHey everyone,
I think SickCity is pretty cool, but I just wanted to highlight another tool that mines Twitter for tweets pertaining to any topic and then applies a sentiment algorithm to it in order to understand how people feel about the product or topic. It was developed by Microsoft's PR agency Waggener Edstrom. Here's the link:
Twendz
Correlation to reality
By bujangaThis could be a great tool. It does however need some correlation to reality. Data collected by the CDC is generally delayed. What Google did was correlate their trend data with the CDC's data. This showed that Google's was a leading indicator for flu trends.
The same thing could be done with Twitter data. I know that their are some surveillance systems in place in various cities that collect flu data in near real time from hospital emergency rooms. This would likely be a good place to look for correlations.
Public Health is not likely to accept data that does not have some proven statistical basis.
re: correlation to reality
By John GeraciYou're absolutely right bujanga. Would you be willing to help guide us in getting some sort of correlation like that in place? That's a crucial next step in the development of the tool.