This blog post was written by a blogger who is no longer at SAS
Wired magazine released the results of its worldwide user survey, which asked users to self-report 3G and EDGE data speeds around the world.
The conclusion is that, when comparing various Apple-branded products, the carrier's network had more to do with data speeds than did the handsets. This sounds like an eminently interesting problem that has relevance to our daily doings. Since I'm an expert JMP user, I do data discovery just to relax at home (sweet tea, data, and exponential regression -- yum!). All I need now is the data.
And guess what? It's publicly available. Using JMP's cut and paste, getting this data into JMP takes only a few seconds. I should have this analysis done before I can finish typing up this blog post.
But, alas, were it only to be that simple. This data is self-reported, and there wasn't much control over what the user could enter as responses. For example, here's a screenshot of a few of the thousands of rows in the data table.
Look at the Carrier column and the creative ways that users entered the company AT&T: "ATT", "AT&T Wireless", "at+t", and so on. There are thousands of rows in this table. It will take me several glasses of sweet tea to go through all these rows and standardize the names of the columns.
Here enters JMP's Recode command. I want to standardize all the flavors of AT&T to have a single label. JMP's Columns > Recode command brings up a window that shows me all the occupants of a column, nicely sorted. Here's the Recode window after another few seconds worth of work, where I have the string "AT&T" on my clipboard and I rapidly descend the column of values using the down-arrow key, pasting the correct value at each step. Easy-peasy:
I do the same with the other wireless carriers, and soon I've got a nice standardized list to analyze. The other self-reported columns can be put in analysis shape with the same command, and soon the analysis is underway -- an analysis, you should note, that I leave as an exercise to the reader. You've got the data, after all.