"No problem. Glad I could help point you in the right direction. Lots of fun stuff you can do with the data and I'm hopeful that some Excel gurus out there will start playing with it and contribute additional ideas."
- AJ Kohn
"Yes @Zack Lim, since I wrote this post the analytics package is now available to everyone and is far more easily accessed.
In addition, you might want to check out the 'Websites' option which allows you to see all Tweets from a website. All it takes is the addition of a meta tag in the head of your site.
Thanks for your kind words and happy analyzing."
- AJ Kohn
"We're not close but I've talked with you enough and watched you long enough to have noticed that you were ... unhappy for some time. But as an acquaintance it's bloody difficult to do anything about that. "Hey Rand, why so serious?" Yeah, that's not going to happen. So I was happy when you stepped down as CEO because you seemed ... lighter. I didn't really care what it meant to Moz. I just knew that someone who I thought was smart and generous had gotten rid of that dark rainy cloud following him around. Like @Will Critchlow seeing you blog about this makes me feel like you're healing. Now, I don't know if I've ever been depressed. I had a breakdown in 2003 after my wife had a miscarriage and being severely unhappy being in the fallback job I took during the dot bomb nuclear winter - literally sobbing uncontrollably at work. And I totally get the mindset of magnifying the negatives and rationalizing away the positives. I did it (and still do to some extent) for years. My employment..."
- AJ Kohn
Not surprisingly, the most important thing is to have the right features: those capturing historical information about the user or ad dominate other types of features. Once we have the right features and the right model (decisions trees plus logistic regression), other factors play small roles (though even small improvements are important at scale). Picking the optimal handling for data freshness, learning rate schema and data sampling improve the model slightly, though much less than adding a high-value feature, or picking the right model to begin with.
- AJ Kohn
These technological advances will enable even better image understanding on our side and the progress is directly transferable to Google products such as photo search, image search, YouTube, self-driving cars, and any place where it is useful to understand what is in an image as well as where things are.
- AJ Kohn
The architecture can solve a wide class of problems from vision, audition, and multi-sensory fusion, and has the potential to revolutionize the computer industry by integrating brain-like capability into devices where computation is constrained by power and speed. These systems can efficiently process high-dimensional, noisy sensory data in real time, while consuming orders of magnitude less power than conventional computer architectures. On one hand, with portable devices: think smart phones, sensor networks, self-driving automobiles, robots, public safety, medical imaging, real-time video analysis, signal processing, olfactory detection, and digital pathology.
- AJ Kohn
"Interesting observations here Andrew. It makes sense that Google would start doing this. The Google My Business profile gives Google geography and category information. From there it's just a hop, skip and a jump to matching it up with query syntax.
The intriguing part is the idea that the highest natural result that has this knowledge graph connection via Google My Business is also given a Knowledge Panel treatment. If that pattern holds then it's a huge advantage for that business."
- AJ Kohn