BLOGS NOW BEING MONITORED FOR *TONE,* NOT TRUTH (ORWELLIAN)

Pasadena Sub Rosa ^ | February 13, 2009 | Wayne Lusvardi

[WayneLusvardi]

You won't believe this one. But, then again, you might. A blogsite called Newstin (is that short for tin news?) is now rating blog posts based on wording *tone,* not truth. This is right out of the novel Fahrenheit 451 or George Orwell's novel 1984. Apparently, any content that is negatively *toned* toward Leftist ideas are labeled *negatively toned.* The web masters at Newstin call their monitoring program *Sentiment Analysis*

Excerpted from Newstin - http://www.newstin.com/tag/us/103990119 Sentiment Analysis is a tool developed by Newstin to indicate the overall tone of an article i.e., whether it has a positive, negative or neutral feel to it. For example, an article about an oil spillage at sea would most likely contain negative wording due to the effects of the incident. Our Sentiment-o-Meter provides readers with an indication of the tone of an article without reading it. Red represents negative tone, green positive and grey is for neutral tone: The higher the scale of the colour (from bottom to top), the greater the degree of the given sentiment.

[Darwin Fish]

80 million gun owners in the US.

Go ahead. Piss us off.

[uncommonsense]

"I'd never heard of this site before this post. How popular and well-known is it? Its impact will only be as great as the number of "eyeballs" it can verify coming to it."

They're new and don't have much traction yet. They supply services to other organizations - possibly governmental agencies. They don't directly take advantage of ad revenue yet and their "listener" technology is functional, but under utilized - still in need of maturity. They've got some talented and famous thought leadership associated with the company. There are a few other companies who provide similar capabilities.

I'm quite intimate with credit bureau capabilities. Besides the deep financial transactional data they have on everyone - anyone that's ever filled out a warranty or rebate card, subscribed to a magazine, had utilities in their name, rented an apartment or owned a home, held a job, etc. - is in their databases. This data can be correlated to health records, court filings, phone and internet access, public records, and other census / demographic data to produce a very complete profile and highly predictable model of preferences / behaviors. It's Orwelian scary, but very useful for business purposes.

I don't personally know anyone who is "off the grid" (no electronic records), so these databases are immense - deep and wide - sometimes carrying 2,000 attributes of data (names, aliases, addresses, products, prices, dates) and hundreds of lines (transactions over time - phone numbers and vendors, employers / position / salary / dates, etc.).