Apple maps (in the loosest sense of the word)

The new Apple iOS6 release has led to a wave of map related hysteria.  There's a terrific Tumblr that's doing a great job of compiling some fantastic carto-gaffes in the new Apple maps app.  My current favorite...

BBC News gives Apple's maps a pretty rough ride as do any number of others including Apple Insider and Gizmodo. It's all very amusing and, one suspects, Google are having a pleasant day at the office having been dropped on iOS6 in favour of Apple's own map app.  But what are we actually seeing?

We're seeing a company renowned for highly polished product pushing out a poor map and that in itself is feeding the sneering hordes.  Sub-standard product would make Steve Jobs turn in his grave but it proves one thing...making a solid, accurate, searchable, well designed and useable online map (and one that works well on a mobile device) is no easy game.  You can't do it quickly and if you don't know what you're doing you're going to make the sort of errors that we're seeing.  Let's not forget though, Google's own initial efforts were ridiculed and they had their own carto-gaffes in the early days...poorly matched satellite imagery, administrative boundaries in the wrong place, labels poorly positioned...not to mention the awful colours, layout and the fact that most of the known world was missing initially.

The problem for Apple is they need to react quickly to get their map app up to speed.  Google have had 7 years of development.  They have learnt a lot in that time, not least that cartography is not easy.  Apple's initial offering seems to suggest they've poured their map data into a really agricultural and restrictive set of styles.  They've done precious little error checking and none of that manual grunt work required to iron out the problems so we're seeing all manner of missing or incorrectly located POIs, weak navigation and search, terrible satellite imagery...an even worse hybrid satellite/map effort and more.  It'll be interesting to see how they improve on their map app over the coming months and years but of course...everyone else has a huge head start and will be working on their own products.

Quite simply, Apple dropped the ball on this one and didn't appreciate the difficulty of making a good map that people can use. With maps and location being absolutely central to how we use mobile devices this is potentially a really big deal. I suspect many will simply not upgrade to iOS6 or will just overcome the mild inconvenience of launching a web browser to hit a better map....or they may just wait until Google publish their own Google iO6 map app in the Apple store which will just make Apple's job even harder.

Me? I use Android...

Zen and the art of cartography

For those that asked and for those that may be vaguely interested, here is my 'Zen and the art of cartography' presentation. I gave this at the British Cartographic Society in June 2012 (UK) and then an updated version at GeoCart 2012 in August (New Zealand). You can view the images below or if you want the accompanying notes, click through to view on Slideshare itself (and select the notes tab below the slides). To be honest...it'll make very little sense without the notes unless you just like pretty pictures!

 

Zen and art of cartography from Kenneth Field

What would Harry have thought of it all?

Today sees yet another bastardization of H. C. Beck's London Underground map. londonist.com report on a new tube map showing Olympic medal winners.

Rarely a day goes by these days without someone or other taking their own dataset, failing to have a single ounce of their own imagination and slapping it on one of the most famous maps in the world. The data often have no connectivity, no relationship with transport or journey planning and absolutely no need to be displayed as a schematic diagram. But hey, if you want to get people interested...imitation is the sincerest form of plagiarism right?

Now even TfL have got in on the act with an "if you can't beat them, join them" attitude for this latest version is an 'official product', available as a limited edition through their online shop. As those who know me will already know...I am an avid fan of Beck's map. I am also a fan of some of the original uses of it as a vehicle for other, re-engineered purposes (e.g. The Great Bear, 1992) but the time has come to say enough is enough. It's also the last straw and I am hereby announcing a forthcoming paper penned by myself and William Cartwright on this very issue...

PS...did I buy it? yes of course...I'm a map addict. Dammit.

Getting steamed up over heat maps

Seen Amazon's Election Heat Map 2012? Here's a screen grab to whet your appetite...

Notice anything strange?  (@briantimoney did, which is where I picked this one up). It's a Heat Map.  A what? I said a 'heat map'.  What the blue blazes is a 'heat map'? I'll tell you what it is...it's a made up name to make a map seem more edgy, more sensationalist, more...sexy. What you're looking at is a sexed up choropleth map. A perfectly good mapping technique, given a new name by Amazon because I guess the term choropleth seemed a little outmoded.  Why not go the whole hog and call it a heat-o-graphic? Actually, why not just give the map a decent title like 'Election reading preferences 2012'.  tell it how it is.  Call a spade a spade and we get it.  Instead, we have to wade through a nonsensical title and then a subtitle "What are American's Reading".

It's along the same lines as 'hot spot map'...that's a map that shows clusters of something. The word cluster being a perfectly good descriptor and, also, having the benefit of being a shorter single word yet hot spot is somehow preferred.  Does it make the map more plausible, more understandable or merely more sensationalist?

Now that the point of this blog has been vented, let's dig a little deeper and see if we can't find some other interesting perspectives...

The map itself shows the distribution of 'political book' purchases from Amazon in the last 30 days, where the books have been categorised as red (Republican Party), blue (Democratic Party) or neutral according to their political reading. Each state's colour is assigned based on the percentage comparison of the top 250 selling blue books compared to the the equivalent red books.  States with a higher percentage of one type over another are coloured in darker shades.

So what does the map show?  It clearly shows a strong leaning towards Republican favoured reading material (oddly for Brits this is a source of constant confusion...Republicans are right-leaning Conservatives, which we designate using blue in the UK!!!). There's a few problems with this.  Books are rarely so overtly Red or Blue so making such a divisive classification adds unfathomable bias into the map.  Does it show a Republican wave taking the US by storm?  Does it show that Republicans tend to read political books more? Does it show that Democrats are buying Republican reading material to figure out what alternative policies and views might be out there? Or maybe it shows that publishers are more likely to either publish right-leaning books or, perhaps, market them more effectively to an audience more willing to consume them? What is clear is that the map does not in any way reflect possible voting patterns, yet by using the same approach as we commonly see on choropleths, sorry, heat maps, of election results we cannot help but make that association.

As a way of seeing what's selling in spatial terms it's quite useful. You can click on each state and see what the top selling political books of each persuasion are.  Here's California's selection which although showing 51% leaning to the red, it's worth noting that California is an avowed Democratic State proving that those who read might not be those who vote:

Now wait just a minute...The Price of Inequality on the left and both The Amateur and Killing Lincoln on the right are each listed twice. That means they are counted twice...that means the assumption has been made that different formats of the book indicate a single independent purchase.  What if people buy multiple editions of the same book? That could easily mean the map shows that Republicans for some reason tend to buy multiple editions of the same book more than Democrats. So we also potentially have double-counting in the data to contend with.

So it's not actually an election map at all, let alone a heat map. In fact, it doesn't even go to the extent of showing us whether 'other' books are far outweighing political reading material which would be interesting in its own right. What we actually have then is a choropleth map showing spurious sales data in its crudest form.  To be fair to Amazon, they don't try and hide what the map shows or how it has been represented. I guess they just decided that a choropleth of sales data isn't a particularly sexy map to get people interested in...