The OECD clearly has a lot of time on its hands – after all, think how long it takes to say The Organisation for Economic Co-operation and Development. But time likewise figures in its research remit, here keying a survey of the time management habits of denizens of its member nations (at least most of them, along with representations from China, India, and South Africa, designated OECD partners), and recapitulated here:
The data – and it’s the Total sheet I’m reviewing – spread out some neat cross-national takes on how respondents apportion their daily round, normalized for each nation’s activities to the day’s 1440 minutes (and I assume normalization means reconciling the inter-national variance in activity definitions; see the Activity category tab). You’ll be interested to know, for example, that South Africans appear to be able to devote 562 minutes a day or night – nine hours and 22 minutes – to sleeping, and that while earmarking but 463 minutes to that circadian necessity we’re left to conclude that the Japanese and Norwegians seem to be drinking the most coffee. It is also noteworthy that respondents from Turkey – the only preponderantly Muslim country in the national roster – spend the most time in religious activity (row 32), at 19 minutes (South Africans come in one minute behind). It’s all interesting, but I can’t resist the geekish aside to the effect that the row 32 numbers are variously positioned – with some middle-aligned, (along the vertical axis), and others right (and you might very well also want to pump up the 8-point text prevailing throughout).
But at the same time the numbers urge a familiar question upon the plenary: For whom is the spreadsheet intended – a public of passive, if interested, readers, or that A-Team of caffeinated deep thinkers who want nothing better than to do something with the data? If you’re siding with the latter insurgents, and want to act upon what it is you’re viewing here with your canteen of slicers (and Slicers) and dicers, then the data – and we’ve seen this before – have to be reshaped. That’s a question we’ve asked in so many words before, and it requires asking here as well.
Let’s see. For starters, we need to do something about row 3, harboring as it does survey date identifiers impersonating as a row of data. And a similar reprisal need be practiced upon those purple subtotal rows, which should be deleted. Leaving them in place and proceeding to drop the data as they stand into a pivot table will do nothing less than double any summed results, and yes, we’ve seen this before (e.g., my January 24, 2013 post).
On the other hand stripping those purple rows will also deprive the sheet of their associated headings, e.g., Paid Work or Study, Unpaid Work. But again, rows of data should be of a piece; insinuating rows of titles into the average daily minutes is tantamount to playing checkers, when in fact we’re playing chess. Moreover, the Men & Wom field in the A column sufficiently identifies the activities subject to the minutes-per-day estimates. But there’s more to be said about this.
But before we expound the larger issue betokened above, swing over to column AC and its swath of computations averaging the numbers lining up to their immediate left. Try cell AC4:
And try explaining it. Note all the references in the expression are alphabetically contiguous; and given its cellular continuity how do we defend the formulation above, when we have
at our disposal? I can’t answer my own question.
And indeed – those averages ignore the three partner nations on the other side of AC, thus plunking the wrong kind of data amid the time-estimate numbers. And the blank AD column has to go.
But there’s that larger issue, reprising an earlier question, the one I asked and attempted to resolve in my August 22, 2013 post. The Total sheet confers field status upon tufts of data that should more rightly conceive of themselves as items relenting to a larger, governing field. Thus, for example, the field-defined member countries in the sheet should be made to submit to the controlling aegis of a Country field. It seems to me, then, that a reconstructed data set would wheel out records looking something like this:
Again, umbrella fields on the order of Gender or Country are far mightier enablers of the kinds of grouping and ordering feats that pivot tables perform, and well justify the necessary surcharge the additional data entry would exact (again, turn back to August 22).
For example, the data reforms I’m (again) proposing would free a pivot tabler to rank national time outlays by any activity, by grabbing the Rank Largest to Smallest marble from the Show Values As bag of alternatives. The point is that Rank Largest to Smallest can do what it does only among items planted in the same field – and not between data settled in different ones.
So if you’re wearing the A-Team colors, why not reorganize the data here and shop it to the OECD – if you have the time?