Text Analytics in the Cloud


politicsbig.jpgIf you're a software developer (I'm not), you might find it easy to throw together a new application. You can handle pulling in data, moving it around, and displaying results—the normal software stuff. If you're analyzing social media content, though, the core functions of text analytics might be something you'd rather not learn. There's an API for that.

These services—some free, most commercial—allow you to skip the R&D and plug text analytics into your system. In a market crowded with me-too platforms, it might be the step that gets you off the entry-level rung of the ladder.

There, that should be enough for another 300 social media analysis startups. :-)

More posts in the "Build or Buy?" series:

Visualization by Thomas Jenkins.

Analytics as a service: AaaS? Wouldn't want to hazard a pronunciation on that one.


You should also include EvoApp on this list :) We combine powerful text analytics in to a one stop portal with social media monitoring for an all inclusive collaboration portal.

We are highly adaptive, and our engine is easily integrated with other platforms. Although relatively new compared to many other companies, our product offers one of the fastest Enterprise Relationship Engines in the market.

Hi Nathan,

Thanks for the list !

My 2 cents: There are three elements people need to have for social media analytics

1- Problem definition. You need to know what it is you want to learn, or at a minimum you need to be able to model the "kind of things" you are trying to learn. Otherwise, you'll end up nowhere (knowwhere ;-) )

2- Clean data source. It's the all garbage in/ garbage out adage. Unless one can produce a close to 100% clean data stream, text analytics is useless.

Clean also mean consistent/relevant. If the conversations that people mine come from diverse audiences or vary between 2 pages deep docs and tweets... then forget it.

3- Algorithms/Technology. In my experience this is max 5% of the solution. P Norvig, from Google, has even proof that when one get enough data, using algorithm/approach A, B or C does not make much of a difference and that the algorithm that performed best with less data, could be the best one with more data.

The technologies that you are listing are great. Building a useful solution with it is ... a challenge. Combining two or more of them to get any kind of result on a noisy data stream. I would not go there.


Comments are now closed for this entry.

About Nathan Gilliatt

  • ng.jpg
  • Voracious learner and explorer. Analyst tracking technologies and markets in intelligence, analytics and social media. Advisor to buyers, sellers and investors. Writing my next book.
  • Principal, Social Target
  • Profile
  • Highlights from the archive


Monthly Archives