ProjectX Blog

Microsoft have release a new version of Virtual Earth Web Service and Map control. There’s a swag of new features including:

• New Mobile Support – Blackberry, Windows Mobile and iPhone
• Internationalised versions – English, German, French, Italian, and Spanish for users in Western Europe.
• New international reverse geocoding – Users can now find international addresses with reverse geocoding, which is now available anywhere Virtual Earth has routing.
• Extended parsing capabilities – Users can expect better match rates for addresses in Australia, New Zealand, Canada, and Puerto Rico.
• Pushpin clustering – The new map control includes the ability to zoom in on a map to better visualize a cluster of points.
• New landmark hints in routing – Customers in the U.S. and Canada can now use maps that feature familiar landmarks, such as gas stations and fast-food restaurants, called out by name.
• New imagery metadata – Users can now find out the relative age of a given image, which will help them assess if the imagery is still relevant to their needs.
• New one-click directions – Through the Web Services, customers can now get directions in one click, choosing from route options by shortest time, shortest distance, or traffic flow.

What’s of interest to ProjectX is the reverse geocoding and landmark directions.

The landmark routing is a step in the right direction. Navigation by landmark is a very common way that we explain directions to people. Its nice to see that Micorsoft is lifting the bar.

We use a lot a reverse geocoding and it’ll be interesting to see how comprehensive it is and how fast the webservice is.

Who’s the coolest company in web mapping ??? Unfortunately its not ProjectX, its Stamen Designs from San Francisco. They have become the company in the map visualisation space. Yesterday, they released a new site – Hurricane maps. Add that to their impressive releases for Trulia, Oakland Crimespotting and mySociety.

Stamen Designs have set the bar for map visualisation. We’re glad that Stamen designs is in the visualisation space and they’re pushing the envelope. It gives us something to aspire to !

Today’s Digital Future Summit 2.0 saw the launch of The National Broadband Map, built with ProjectX technology. Its purpose is to help improve access to broadband, especially in the regions, through “demand aggregation”: pooling the demand for telecommunications services in a specific geographic location.

Individual users might find it hard to convince a telco to roll out services in their area, but when those users combine, they can often reach the critical mass required to attract investment in the provision of new services. For instance, if you’re a graphic design business in a small provincial town, you might struggle to convince a provider to extend their infrastructure to reach you; but if there is also a school, medical centre and library down your street, plus several other businesses and private residents who are also interested, then the telcos might sit up and take notice.

To this end, the map shows existing state sector locations, as well as allowing businesses and private residences to enter their own locations and describe their level of demand. It also shows existing network infrastructure, and includes all the address search, autocomplete, zooming and panning functionality that is familiar to ZoomIn users.

The system was developed by ProjectX together with the State Services Commission, and involved the collation, geocoding and checking of thousands of state sector locations, and the conversion of network data from all sorts of formats and projections, as well as building on existing ZoomIn Mapping System API functionality. For instance, the ZMS has always had the ability to display vector layers (polylines and polygons) as overlays on the maps, but the size and complexity of the network infrastructures required some hard work from our developers to optimise the performance of these layers, as well as dealing with the inevitable curly cross-browser compatibility issues.

In keeping with the Web 2.0 theme, the Map is labelled “Beta”, and it will evolve over time. Partly that will be due to the growing information base, from user-added places and as we receive more state sector locations and existing network data. There will also be functionality and usability improvements over time, and we encourage user feedback so that we can make this as comprehensive and easy-to-use as possible.

Following up from yesterday’s post about the Open Source Awards, here’s a new local discussion group about the use of open source software in Geographic Information Systems (GIS). As a web mapping company, we make heavy use of open source software in both the web and the mapping sides of our business: I’ll come back to the web side, but here are some quick notes on the open source GIS packages that we use.

• PostGIS: I won’t make great claims for its performance, but the spatial extensions to PostgreSQL are pretty impressive and getting more powerful all the time.
• MapServer: this hasn’t been our first choice for map tile rendering in most cases, but now that version 5.0 is out with its gorgeous anti-aliasing (thanks to the AGG library) it’s looking more and more appealing.
• QuantumGIS: it’s not exactly ESRI or MapInfo, but it’s quite useful as a basic desktop GIS app, and the ability to export to SVG (for further tweaking and polishing in Inkscape, or for XML hacking) is great for publication work.
• OGR utilities: indispensable! Good old-fashioned simple command-line tools for exploring, converting and generally mucking around with the obscure world of GIS data formats.

We like to think that our technology is being used solely for good rather than evil, but there are some applications where it’s hard to tell. Stalking would be a no-no, but celebrity stalking doesn’t seem so bad. That’s the purpose of the Wellingtonista Celeb Vista group on ZoomIn, which is fully explained in this post on the Wellingtonista.

Simply put, it’s a group for tracking celebrity sightings, but Wellington being Wellington, the celebs being stalked are somewhat more highbrow than on other celebrity-stalking mashups, hence the inclusion of Peter McLeavey along with Bret McKenzie and Shortland Streeters. At least there are no politicians … yet.

IntensCITY week, the celebration of Wellington’s urban spaces, is just about to kick off. A key part of the week is the INSite exhibition, which consists of eight shipping containers distributed around the city as artists’ spaces and installations. Here’s a ZoomIn group to show their locations (or you can go directly to the group map on ZoomIn).

As these are ZoomIn places, you can of course add your own photos and comments and edit the descriptions. I’ve kicked that process off, but if you manage to get some good photos or have some responses you want to share, please go ahead and do so.

One of the things that we’re always trying to do with data visualisation is show more variables, while trying at the same time not to clutter the map and overwhelm the viewer. When trying to show more than one variable across a map, it rapidly becomes impossible to grasp the information in one go. Edward Tufte, on page 153 of his classic book “The Visual Display of Quantitative Information”, describes one such attempt to map two variables through interaction of two colour schemes as “a puzzle graphic”, “experienced verbally, not visually”.

But that doesn’t mean that it’s not worth the effort. The distribution and interaction of two variables across a region is inherently complex, and to expect every nuance to immediately leap out at a viewer without reflection and complex analysis is unrealistic. My experience of reading professional weather maps that show multiple overlapping contours (e.g. of pressure and temperature) leads me to believe that any multivariate geographical visualisation will require some practice to get the most out of: to learn that where the isobars and isotherms are perpendicular, thermal advection is occurring.

If two variables are too much for Tufte, then the map below would appal him: I’m experimenting with showing three census variables at once on a map of central Wellington. I’ve mapped residential population density to the red channel, density of office workers to the blue channel, and density of other workers (e.g. manufacturing, retail) to green.

It’s certainly complicated! But nevertheless, certain patterns do immediately leap out, and the more I look, the more I see. Purely residential neighbourhoods are quite distinctive in shades of pure red, and variations of population density can be seen quite clearly. The CBD stands out very plainly, but on second glance there’s also a difference between the Lambton Quarter and Thorndon: the cyans of the former show that office workers are balanced by others (probably retail), whereas the latter is much more of an office ghetto. Te Aro is quite a patchwork of greens and earthy tones, showing that a growing residential population is starting to complement the retail, entertainment and tradition light industrial uses of that part of town.

So, while bearing in mind Tufte’s warnings, I think that this sort of trivariate thematic mapping might have a lot to offer. In particular:

• High data density: three separate variables across densely measured meshblocks.
• Emergent patterns: without having to explicitly code “mixed use districts” or “CBDs”, they emerge from the interaction of two or more variables.
• A meaningful grid: census meshblocks are mostly based upon city blocks, so variations in urban form, such as the grid pattern of Te Aro versus the curvy hill suburbs, are easily visible. The familiar forms (at least to locals!) of the wharves guide the eye without explicit use of contextual layers.
• Exploiting human visual processes. While I’d concede Tufte’s point that one has to consciously remind oneself of what certain colours mean, the fact that the RGB system has a neural reality makes it easier than a more arbitrary system would have done.
• Metaphorical power: I deliberately chose to make light colours represent high density and black represent an absence of people, which is the inverse of the usual approach. That gives a map full of neon colours, evocative of “city lights”, and fits with my own preference for dense and vibrant cities. It also brings to mind a mosaic, with its connotations of colourful diversity. As always, symbology reveals ideology!
• Visual appeal. Well, at least I like to think it looks good! While it wasn’t deliberate or expected, I think it’s ended up quite reminiscent of early Paul Klee.
• Depth of exploration. While the two previous points may seem relatively trivial, a map that is attractive and evocative is more likely to bring people back to explore further than an ugly or dull map would. Every time I look, I keep finding more patterns and intriguing anomalies, and it keeps raising new questions to ask of the data.

While I may be a fan of the “intuitive” and “idiosyncratic” nature of neogeography, we all recognise the importance of getting the core geographic information right. We’re about to start a comprehensive update of all our mapping tiles, and while we were able to get out a relatively quick fix for the Wellington bypass, it will take a bit of effort to do it properly for the whole country.

While we pay good money for our base data, it’s a fact of life that nothing’s perfect, and the physical world sometimes moves too fast for surveyors, councils and data providers to keep up with. We already have a list of updates and corrections that we’ll probably have to make manually, and we’ve had useful reports from ZoomIn users that we’ve added to the list. But we want to ask you, our blog readers, whether you’ve noticed any errors, out-of-date areas or places that could do with a tweak.

Along with that, I’d like to know if there are any improvements we could make to the appearance of our maps. As just one example, as a supporter of pedestrian-friendly cities it has always seemed odd to me that pedestrian streets like Cuba Mall have the same symbology as dirt tracks through the bush. There’s a field in the raw data that distinguishes “malls” from walkways, and if I get the chance I’d like to represent the difference visually. Are there any visual quirks or infelicities that you’d like us to take a look at?

We can’t guarantee that we’ll get every suggestion into the upcoming map release, but we’re keen to get your feedback on what would make our maps the best they can be.