That chart above is what political disaster looks like as reflected in the blog community. Specifically it tracks the volume and tenor of the posts about Eliot Spitzer, the now-former governor of New York, in the days immediately following his outing at “Client No. 9”, the Emperor’s Club VIP customer who met with a prostitute in Washington’s Mayflower Hotel.

Relevant Noise, the blog-mining service that produced the chart, found that Internet posts concerning Spitzer rose about 4,700% after the news of his possible involvement in a sex scandal broke on March 10. In the week prior to the story’s break, there were only 104 posts in the blogosphere concerning New York’s top executive. But in the three days starting with the first headlines, going through Spitzer’s quick admission of guilt to his resignation, more than 4,980 posts cropped up discussing the governor’s predicament.

RelevantNoise also took a look at the tone of the posts in these conversations and determined that on March 12, the day Spitzer formally resigned office, 66% of the conversation was negative toward him and 6% linking him to the words “hypocrisy” or “hypocrite”.

The chart is an interesting time-lapse snapshot of how an issue can quickly come to dominate the news—and then settle back into a kind of low-level hum in the national chat.

Relevant Noise examines blog traffic, automatically examining content from a healthy selection of the 80 million blogs in the world today along with forums, message boards and other consumer-generated content. The company, a division of digital marketing agency Zeta interactive, uses data mining technology, natural language processing and “tone extraction” technologies to determine the volume and tone of what’s being said in social media about key issues.

The tool is primarily aimed at corporations who want to keep tabs on what bloggers and consumer commenters are saying to each other about their brands and products. But this presidential election year is providing a lot of opportunity to prove the concept of social monitoring by putting RelevantNoise to work on the political blogs.

For example, of the bloggers who expressed a preference to see one of the current candidates voted into the white House this November, Barack Obama won 57% of those explicit endorsements. Clinton was the choice of 28% of blog posts in the last 30 days—or about half the appeal of her counterpart for the Democratic nod– while John McCain trailed with 15% of the blogger vote during the same period.

Of course, there is a school of analysis that suggests bloggers are more likely to lean to the Democrats, so the conversation in social media may differ widely from that in the voting public at large. Only November will tell.

From the demographic perspective, Clinton got a larger amount of support from female bloggers during the month than either of her rivals: 61% of her blogging partisans were women, compared to the majority of males in favor of both McCain (66%) and Obama (58%).

“There’s a lot of need to understand what people are thinking about these candidates,” says Al DiGuido, CEO of Zeta interactive. “RelevantNoise has the ability to track what people are saying in social media on a day-by-day, hour-by-hour basis and then to aggregate it and examine the tone around these candidates. What words are being used to describe these people?”

DiGuido says that RelevantNoise’s efforts to measure the “tone” of mentions in blogs is actually a very precise algorithm built around sentence structure. The result is a tool that, he claims, can reliably filter out “splogs” (spam-laden blogs that exist only to promote other Web sites) and flames, using metrics such as the number of exclamation points in a post, to leave only the most authentic content.

Last December the company used that tool to examine the boost in both traffic and tone that two candidates got from their respective celebrity endorsements. Interestingly, RelevantNoise’s research found that Obama’s support from Oprah Winfrey did little to boost either his traffic or his positive buzz, at least among the blogging electorate. On the other hand, former candidate Mike Huckabee got a 66% bump in social media mentions from lining up the support of action stud and fitness pitchman Chuck Norris. But he also got bumped down in his positive buzz, from 82% positive before the Norris announcement to 77% after.