I'm learning R. It's brilliant. End of.
Ok not quite end-of. I tried Excel and SPSS first. I love Excel. In fact we have a long-standing co-dependant relationship. But it is time for cold turkey and here is why.
With all the hoopla about tablets and touchscreens I am suddenly back thinking about plain-old-typing. Sure it is slow, error prone and requires a large surface area to run your hands over to do it any way efficiently, but there are some things that GUI-driven modes are just terrible for. Working with data is one of them.
I have a big Excel file with 25,315 rows and about 27 columns. I am doing lots of stuff with it like averaging, summing, counting, filtering etc and matter how I layout the file, including using various tables, new tabs etc. sooner or later I find myself faced with...
The scroll of deathTM
There are neat tricks around this in Excel for instance you can select (highlight) a swathe of spreadsheet by using calliug Ctrl-G, typing the reference of the cell you wish to extend the selection to, then pressing shift and then Enter. TA-DA! This allows you for instance to select thousands of specific cells anywhere in your spreadsheet where dragging using the mouse would have taken days if not minutes.
But eventually you end up scrolling in a spreadsheet and you end up with some of your data in one place and the rest in another. It's horrible because scrolling is painful. Scrolling is not a good solution to anything.
Enter R. With a program like R the metaphor is different. There are no scroll bars. You cannot organise your data as you would in Excel, where the temptation is always to be able to "see" as much data as possible. Rather you don't care where the data is and you summon it as you need it. These incantations of the data are nor performed with palaeolithic scrollbars or mouse pointers. Instead we use the neolithic earthcrushing, shapeshifting technology that is the keyboard. The mighty artefact that puts the digit back into digital. Long live keys.
0 comments:
Post a Comment