Amplitude - All Comments

re: Measuring ROI on Semantics

SteveS — Tue, 09 Mar 2010 16:33:58 GMT

I agree with Balaji... I think tracking the viral nature of the social networks is key, but sentment analysis of these results is still the real killer app.

re: Measuring ROI on Semantics

balajik — Tue, 09 Mar 2010 15:31:28 GMT

In this era social web ,I feel the following would be a good starting point.

1)Number of tweets for a that url.

2)How many people share that url on their social networks.

re: Your Sick Isn't My Sick

Alexandra Stalnacke — Mon, 08 Mar 2010 14:39:57 GMT

Being a swede, I say this almost with tears in my eyes: Mike, it's not just your SAAB that's sick. SAAB's sick - sick in the most negative way.

SAAB used to make us proud, but then, what happened?

re: Your Sick Isn't My Sick

SteveS — Thu, 04 Mar 2010 15:33:11 GMT

Oh, it's a sick car all right... and I mean it in the best possible way! :)

re: Your Sick Isn't My Sick

mikepetit — Tue, 02 Mar 2010 19:47:35 GMT

By the way, that photo on the home page (good find, Dave) is the same model and color, etc., although not a shot of my car. They only made 300 of them; not too many left. That car is *sick*, no? :-)

re: Why this stuff is hard :-)

sderose — Thu, 10 Dec 2009 20:06:05 GMT

Well, genre ID would be a good first step, and fairly tractable. Though here we want a mixture of genre (review, tutorial, discussion group, technical analysis,...) and domain (music, sports, cooking,...).

As long as your dictionary has information associated with specific word-senses or registers, and you have some information about the domain and register in use, the word-senses can be picked out, probably pretty reliably.

re: Why this stuff is hard :-)

Alexandra Stalnacke — Wed, 25 Nov 2009 09:05:03 GMT

Hi guys! Yes! interesting indeed.

Consider a really green lawn; a ceiling so moist that it's turning green; so called Green programming; or a really nice and moist chocolate cake compared to the embarrassment of moist hands...

re: Why this stuff is hard :-)

mikepetit — Tue, 24 Nov 2009 20:02:18 GMT

Very interesting comment, Steve. How would we implement such a genre-specific scheme? Permit people to volunteer some kind of genre code in the API?

re: Why this stuff is hard :-)

sderose — Tue, 24 Nov 2009 16:43:21 GMT

Actually, we don't do too bad. Top topics come out as Entertainment, Music, and drummer. The Polarity on the drummer comes out negative, which is wrong but not surprising.

Reminds me of a gig my sister played when she was still in high school. Guy came up to her afterwards and said "girl, you play really *bad*" -- and she was all upset until somebody eventually explained the idiom to her.

I wonder how feasible it would be to handle genre-specific jargon like that -- say, treating "bad" as positive in discussions of music. Do people ever actually say "bad" in music reviews? Not much, I suspect; critics tend to find fancier ways of panning people ("full-time consideration of another endeavor, might be in order").

Which reminds me of a great book: "The Lexicon of Musical Invective". A huge collection of really scathing reviews....

re: Online Advertising Can Make Us All Cool

SteveS — Mon, 19 Oct 2009 13:23:04 GMT

What a very apt OpenAmp result set! Hit the nail on the head. And by the way, the yellow Saab is SWEET!

re: Online Advertising Can Make Us All Cool

balajik — Fri, 16 Oct 2009 20:01:49 GMT

Yeah, I just tried Amp'ing the above text . The following is the output from OpenAmplify:

* Business

o Advertising: 52.000000

+ ad:39.000000

# Polarity: Positive ,0.306155

+ advertiser:13.000000

# Polarity: Positive ,0.904886

* Automotive and boating

o Car: 37.000000

+ Saab:25.000000

# Polarity: Positive ,0.808858

+ Lexus:12.000000

# Polarity: Negative ,-0.076049

Look at the Polarity for Lexus and Saab . Doesn't that sound Interesting.?

It is cooool.

re: Eight Years Later

SteveS — Tue, 15 Sep 2009 14:03:55 GMT

I was working in Georgetown, right across the water from the Pentagon. I had heard that a plane had hit the World Trade Center, but thought that a small aircraft had just gotten too close. The gravity of the situation wasn't apparent to me until the woman in front of me in a coffee line turned to me and told me that the Pentagon was just hit. When I finally realized that we were actually under attack, I went into a mild state of shock... how could this happen, and on such a beautiful day? I still can't believe it now, 8 years later...

re: Eight Years Later

ramS — Mon, 14 Sep 2009 21:13:30 GMT

I remember watching the towers burn on CNN. My wife and one year old son were in an office building ~100 yards from the White House so I was pretty worried about more planes hitting Washington. Fortunately, the day care evacuated the building in an orderly fashion and my wife was able to find them outside. The metro was shut down, but she had the stroller with her so they managed to walk out of downtown DC and eventually catch a bus back to the house. Cell phone service was very intermittent, so I didn't know that they had gotten home for several hours.

I was working in Virginia and they were advising that all the bridges were closed and the evacuation routes (ie the Beltway) were totally jammed, so I waited until the afternoon to drive back. I took some back roads and went over Chain Bridge without too many problems. It was a clear sunny fall day, I took my son to the local park and he played, totally unconcerned, which was very heartening to watch.

re: Eight Years Later

DaveW — Mon, 14 Sep 2009 17:44:38 GMT

Mike,

I am glad you wrote down your memories publicly. This kind of thing builds an important collaborative memory.

I took the opportunity you brought forward from this post to write down some of my own memories:

community.openamplify.com/.../eight-years-later-my-story-dave-weinberg.aspx

re: Cloud Computing Works!

ramS — Thu, 10 Sep 2009 13:53:41 GMT

Cloud Computing works, but a little weather forecasting always helps. How do you plan to use your daily allotment of OpenAmplify transactions?