This morning MapQuest launched their US support of OpenStreetMap at open.mapquest.com. In playing with the interface, I noticed how MapQuest added a tab at some point for showing and hiding the sidebar of search results and other associated design choices and differences.

MapQuest uses an “Outie” tab (highlighted in the screenshot below). The design choice was clearly to make it very explicit for users to show and hide the sidebar as it protrudes into the map interface. The pan and zoom controls are on the right-hand side, so when you toggle the sidebar, the controls stay in the same location. Another interesting aspect is how the map resizes. In MapQuest, the same geographic center and extents remain in the screen center – so as the sidebar closes the map shifts to the left and expands slightly.

Curious about how this varies, I checked in Google Maps. They chose to be much more subtle about their sidebar toggle. It is an “innie” that is subtly hidden within the header. Closing the sidebar turns the selection to an “outie”, but still remains out of the way in the header. A particularly interesting decision is that the map remains in the same location – so the zoom pan controls move but new areas of the map are exposed. So while the user doesn’t have a context shift (points on the map remain in the same area of the screen) the map now needs to be recentered so that the focus area can be kept in the center.

Lastly, looking at Bing maps it’s a bit of a hybrid between the two. The sidebar tab is in the header like Google, but hiding the sidebar re-centers the map like MapQuest. The controls in Bing are in the header, so they don’t need to shift when the sidebar is toggled. What’s perhaps a little confusing is there is also an “X” close button next to the sidebar tab that clears the search results. It’s not really clear why you would want to clear results – and instead there should be an option to go back to the “table of contents” or similar concept that shows simple links for directions and more.

Much like the emergence of Pan-Zoom bars have become the defacto standard in web mapping interfaces – the sidebar has also become nearly ubiquitous. So it’s interesting to see the slight variations as interaction designers experiment with what users will find easy to understand.

This week on the DC National Mall there is the 2009 Solar Decathlon. It’s a contest between 20 student groups from around the world that build, on the mall, sustainable, energy efficient, and modern houses. The competition measures their efficiency, quality, resource usage, and design. It’s a one week miny village.

So of course, like any village, it needed to be mapped. I went down Saturday afternoon and captured the locations and names of all the buildings and paths that will be up for the week. These are then loaded into OpenStreetMap with start_date and end_date tags that notify the renderer when the features should be visible. It’s a similar model to how Burning Man is mapped year after year as it walks along the Black Rock Desert.

It’s ephemeral mapping – objects that exist in real place, but just for small slices of time. Important as any other building, yet typically relegated to flyers or verbal descriptions.

The fascinating part of projects like this is that OpenStreetMap allowed me to create a map that was useful and immediate. Within minutes of uploading the data, it was available as rendered tiles, vector data, and downloadable to GPS units and iPhones. People on the mall could immediately view the local map with this new information.

It’s a nice demonstration of how community projects like OpenStreetMap will continue to innovate faster, and more openly, then other ‘crowd-sourcing’ options.

I was lucky enough to be invited to enjoy the US 2008 election night at NPR main headquarters in Washington, DC.

Of course, within minutes of walking in I started getting feature requests from reporters and bloggers. One of these was the ability to easy page through reports by state (hint: http://votereport.us/?count=200&state=VA).

We did our best to get a quick web-based map visualization up that would be usable by a large number of people with basic browsers. This limited the number of markers to 200 (one reason we chose to use Flash for our current rendering tool). However, one way of addressing this was to offer a KML file that works very well in GoogleEarth for large sets of markers. Here are the 10,000 reports as of this evening.

Another feature we snuck in recently are some simple statistics on the number and time of reports today at http://votereport.us/reports/stats:

To close out this post – we searched the database to pull up this great audio report by keema: http://votereport.us/reports/9240 – I highly suggest you listen to it.

Like a good geo geek I spent my “holiday” digging fast and deep to contribute to the community around providing information assistance and monitoring around Gustav’s path through Cuba, the Dominican Republic, and Louisiana. The result was a quick prototype map: http://crescentmaps.org/gustav/.

There were several issues at hand – one was the lack of available, accessible data. I started adding data sources to the wiki page, and when available uploaded them to GeoCommons Finder! tagged ‘gustav’. Sean offers great reasons on the importance of data sharing (and some more really great previews of Maker!).

Ed shares his thoughts on the overall experience and also questions the impact? Having spoken often with people such as Jesse Robbins and Mikel on their work in helping out in disaster response technology development and deployment the primary lesson I’ve picked up is: iterate.

Of course, Gustav fading out does not mean everything is clear and over. New Orlean citizens are still not supposed to come back into the city, there are still major power outages, and there was already a huge amount of work continuing from the Katrina recovery. There has been a marked neglect of public assistance on non-US regions such as Cuba and the DR. In addition, there are more storms coming up this season – so make sure and pitch in to help out if you can.

Erik Wilde was pointing out the disparities between Google Maps and Google Ditu, or their Chinese version of maps. However, Google Ditu doesn’t have satellite imagery.

There are several easy ways to fix this. The first was to look at the Ditu tiles, and confirm they are the same as Google’s nominal tiling scheme. Which means you can add the China Street tiles as a simple GTileLayerOverlay with Google Maps standard satellite view underneath. This was incredibly easy with Mapstraction and I put up a demo here.

For bonus points I even added a Mapufacture syndicated feed of Erik’s venues for LocWeb2008 and nearby Wikipedia articles from Geonames.

The other way

The terms of how mixing Google’s various tiles together isn’t clear. So the other way to address his issue is to use the freely available data.

Namely, OpenStreetMap for roads, OpenAerialMap or other remote imagery, and run in OpenLayers. Here is the same map done with open data and open source. The resolution or completeness isn’t there yet, but you can see where it’s going and the ability to be use the information as you want is very appealing.