Manovich's ideas are certainly interesting, though I'm not quite sure I entirely understand them. In some sense there's an argument for the expansion of a software type that is culturally aware by tracing many types of data, but to me this seems like artifice. And given that upon the construction of such a system, it would be rife for misrepresentation, as happens so often on Facebook due to the manipulation of its analytics system, I'm not sure I entirely understand the value of creating such a system. His idea seems to be that by creating this new data we'll have a better understanding of older data, but it seems more like we're shifting the focus of one set of data onto another or possibly many, further complicating rather than providing a better overview of cultural information.
He also claims that it will provide an expanded understanding of data, which is to say we will have more data to look at in a qualitative sense, but I think what he means is that the new set of quantitative data is more like a tool which could potentially be useful for qualitative analysis. While tracking a player in a MMO would be interesting, I'm not sure how tracking such interactions isn't quantitative. It seems that his idea of a qualitative analysis is something akin to tracking multiple pieces of quantitative data. I tend to see qualitative information as the reasoning rather than the rhyme.
Sketch
My sketch is attempting to look at death in particular, and finds some interesting points where, every 2-4 months, there's a sudden uptick in the searches on it, and around Christmas also seems to be a popular time to search about death. While it's an interesting topic, the searches are still not quite specific enough yet, and these tick points, where a bunch of books about death are suddenly checked out, need to be examined more closely. Still, the data seems interesting so far, particularly in searches related to historical death and poetic death.
http://www.mediafire.com/?h3b75k278kztqsw
MySQL for visualizing and CSV provided below.
http://www.mediafire.com/?jsal6nrtnophier (CSV)
SELECT DATE_FORMAT(ckoutdatetime, '%Y-%m-%d'), count(barcode) FROM transactionsall
WHERE title LIKE '%death%' and subject1 LIKE '%death%'
GROUP BY DATE_FORMAT(ckoutdatetime, '%Y-%m-%d');
A more advanced SQL query which looks at a specific class of death (namely historical and poetic - this focuses on death as a historical/literary event)
SELECT DATE_FORMAT(ckoutdatetime, '%Y-%m-%d'), count(barcode) FROM transactionsall
WHERE deweyclass > 800 AND deweyClass < 1000 AND title LIKE '%death%' and subject1 LIKE '%death%'
GROUP BY DATE_FORMAT(ckoutdatetime, '%Y-%m-%d');