[Another in an ongoing series of posts related to But Still They Blog: The Liblog Landscape 2007-2009, a project currently in preparation.]
I’m stepping back from this project for a few days, as soon as I write this post–for two good reasons, which I’ll get to in a moment. Stepping back does not mean abandoning, not even close: It means that I won’t do any work on the project for a while.
The second time around
When I wrote The Liblog Landscape 2007-2008 (still available, and it would be lovely to get from the current total of 56 copies sold to, say, 100), I spent a fair amount of time experimenting, rethinking, redoing, futzing around with, and finally deciding on plausible ways to show the metrics and what they might mean. Some of that time was spent in the usual spreadsheet screwups the first time around (what? you’ve never screwed up a spreadsheet? really?); some was spent trying to understand what embedding an Excel spreadsheet in the Word document that includes a graph from that sheet really means (otherwise known as “why there are a series of small partial spreadsheets in addition to the big primary one”).
I wound up with 65 tables and 16 graphs (line graphs and scatter plots), in addition to the somewhat-less-than-607 tables in the individual liblog profiles (a few blogs didn’t have enough metrics to make a table).
The second time around, I understood the metrics better and I scrapped some of them (anything having to do with use of illustrations in blogs, subgroup identifications based on country or affiliation of blogger). The spreadsheet came together a little more rapidly, and plugging quintiles back into individual liblog profile tables was much cleaner and faster than before. (If some of this sounds like gobbledygook, go back and read earlier posts or buy the book: “Quintiles” are divisions of the studied universe into five groups, e.g., the 20% with the most posts, the 20% below that, the middle 20%, the lower 20% and the lowest 20%); “liblog profiles” are individual writeups for individual blogs, including a table showing a group of key metrics. In the existing book, all the profiles come at the end, in a 147-page final chapter.)
That also meant that putting together the metrics-based chapters, or at least the first five of them, was fairly rapid–and I say “was” advisedly, since I just finished the draft Chapter 5 yesterday. This time around, instead of putting all the profiles into one huge lump at the end, I’m including them in other chapters, usually the first time they’re mentioned within a chapter. So, for example, Chapter 1 (which, along with an introduction and notes about choice of software platform, discusses the age of blogs) includes profiles for 87 “pioneers”–blogs that started before 2004. (Actually, 86 pioneers; I use this blog as an example to explain what’s in a profile). Chapter 2, on overall posting frequency, profiles 56 blogs that are among the most frequent in 2009 and weren’t already profiled… and so on.
So here I am, with five chapters (169 book pages, 41,000 words, 338 profiles used so far of 521 total). That covers all of the base metrics, but not combined patterns of change, correlations and chapters that will be less grounded in metrics.
But…well, the problem with it being faster and easier is that I’m not sure I’ve seen what’s there to see in the results. I may be too close to the numbers at this point.
And, to be sure, even though the final Cites & Insights for 2009 could be as much as six weeks away (but is more likely three to five weeks away), I should probably get started on one or two essays for that issue. (The next due date for my remaining other column isn’t until December, so I won’t work on it until November.)
Those are the two reasons for stepping back for a bit:
- To work on “something else”–although it could be related–that will form part of C&I and break away from this for a little while.
- To be able to return with a fresh look at what I’ve done so far, and possibly (probably?) beef up some of the chapters with new insights on the data…and, maybe, to modify the remaining outline based on what I see or don’t see.
(For example: Maybe it would make sense to look at a subset of the 521 blogs that might be called the “common blogs”–ones that have a significant number of posts in all three years, ones that have full metrics for all three years, ones that aren’t current awareness services in blog form–and see whether those blogs, possibly 200-300, show more distinct patterns than the overall set.)
So it’s not all process
To reward anyone who’s made it this far (anyone?), a couple of notes about the universe of 521 blogs in this year’s study:
- I was expecting fewer new blogs from 2008 than there had been from 2007 (which in the earlier study was down about 22% from 2006 and 35% from 2005), but not as many fewer: This study includes a mere 11 blogs that began in 2008, down from 103 in 2007. Either newer bloggers aren’t bothering to add their blogs to LISWiki or it’s getting tougher to gain enough readership and links for a reasonable Google Page Rank–or there just aren’t as many new and sustaining blogs.
- Oddly enough, for this smaller universe, some earlier years are much closer together: Where The Liblog Landscape 2007-2008 includes 160 blogs that started in 2005, 144 that started in 2006 and 104 that started in 2007, deletions and additions are such that the new universe includes 127 starting in 2005, 123 starting in 2006–and 103 starting in 2007.
- Checking for the most recent post on September 30, 2009 (a scan completed in one long day), 42% of the blogs had a post within the week, 52% within the fortnight and 63% within the month–but only 84% had posts within six months and 90% within a year.
Now I’ll set the spreadsheets and Word chapters aside and go print and arrange lead sheets for source material for one or more of three likely essays for the December 2009 Cites & Insights, the preparatory step to writing the essays…