Webmonkey » Location

Mining Flickr to Build 3D Models of the World

Scott Gilbertson — Tue, 30 Nov 2010 18:04:32 +0000

Microsoft’s PhotoSynth tool is jaw-droppingly awesome. But, because it’s a Microsoft project, the technology is unlikely to appear on some of your favorite non-Microsoft online apps, like Google Maps or Flickr.

However, our friends at ReadWriteWeb stumbled across a very similar tool — at least in terms of the end result — developed by the University of North Carolina in conjunction with Swiss university, ETH-Zurich.

The team has developed a method for creating 3D models by pulling in millions of photographs from Flickr and using some fancy algorithms to generate 3D models of local landmarks. Perhaps even more impressive the results can be generated using a single computer in under a day.

Project lead Jan-Michael Frahm touts the project’s efficiency saying, “our technique would be the equivalent of processing a stack of photos as high as the 828-meter Dubai Towers, using a single PC, versus the next best technique, which is the equivalent of processing a stack of photos 42 meters tall — as high as the ceiling of Notre Dame — using 62 PCs. This efficiency is essential if one is to fully utilize the billions of user-provided images continuously being uploaded to the internet.”

While the results are cool and would make an impressive addition to any number of geo-based services, more serious use cases include helping disaster workers get a better idea of where they’re headed and the extent of damage.

So far the researchers have released a movies demonstrating the technique on landmarks in both Rome (get it? built in a day…) and Berlin, and the results are impressive. For more information on how the process works, check out the UNC website.

See Also:

Personalize Your Map With a Custom Map Marker

Michael Calore — Thu, 07 Oct 2010 17:35:51 +0000

If you’re adding a map to your website, why settle for the vanilla design when you can customize it and leave your own personal mark?

This tutorial will show you how to create a custom map from scratch, then add a little unique flavor to it by replacing the standard “map pin” icon with a custom icon of your own design.

To do this, we’ll be using Mapstraction, a library that creates map code that can be reused across all the big mapping providers (Yahoo, Google, et al). Mapstraction also allows for multiple types of customization such as custom info bubbles and graphics like the one we’ll be dropping onto the map.

Note: This tutorial is adapted from the book Map Scripting 101 by Adam DuVander. Adam is a former Webmonkey contributor and executive editor of Programmable Web. In his book, he shows how to use all of the features of the most popular mapping APIs, and how to mash them up with data from other sources like events calendars, weather services and restaurant review sites to make a variety of custom maps.

This exercise comes from chapters 1 and 2 of Adam’s book, and it is reprinted here with his permission and that of the book’s publisher, No Starch Press. It isn’t a word-for-word excerpt. It has been slightly adapted to work as a web tutorial. You’ll find dozens of in-depth exercises — including the full version of this one — in the book itself.

Create a Mapstraction map

Mapstraction is a little different from Google Maps and Yahoo Maps. Mapstraction is an open source JavaScript library that ties into other mapping APIs. If you use Mapstraction, you can switch from one type of map to another with very little work, as opposed to rewriting your code completely.

Using Mapstraction limits your risk to changes being made to an API. For example, if your site’s traffic takes you beyond the limit for your chosen provider, or the provider begins placing ads on the map, Mapstraction lets you switch providers quickly and inexpensively.

To use Mapstraction, you must first choose a provider. In this example, I’m using Mapstraction to create a Google Map.

Open a new HTML file and type the following:

 
   
    Basic Mapstraction Map

Just like you would for a normal Google Map, we include Google’s JavaScript (line 4).

For this code to work, you also need to download the Mapstraction files. Go to mapstraction.com or the project’s github page, and follow the instructions to save the files in the same directory as your HTML file. Best practices would dictate that you keep JavaScript files in their own directory, separate from your HTML, but I’m simplifying things for this example.

The Mapstraction files you should have, at minimum, are mxn.js, mxn.core.js and googlev3.core.js. You may also have files for other providers, such as yahoo.core.js. The only one we need to reference in our HTML code is mxn.js, which loads the other files that it needs, including those we pass it in the file name. Then, in the create_map function, we let it know which type of map we are creating.

Once you have your Mapstraction map, save your HTML file and load it in a browser. The result should look exactly like this.

This Google map, created via Mapstraction, should be centered on No Starch Press’s neighborhood in San Francisco.

As you can see, the HTML hooks are minimal. Some styling to determine the size of the map and an empty div tag with an id attribute are all that’s required. The JavaScript function create_map() takes over and makes calls to the API. This function can have any name you want.

The minimum amount of information needed within the create_map() function is a map type (googlev3), a center point (using a latitude/longitude pair) and a zoom level (Mapstraction’s tightest zoom level is 16, so I backed off one notch to 15, about six blocks across). Then, we pass those options and reference the div tag’s id to create a map.

Add a marker to your map

To add a simple marker to your map, you just need to use two Mapstraction functions. First, create the marker. Next, add it to the map. The reason for these two distinct steps will become clear in further projects when we start to use advanced options, such as custom marker icons.

Let’s see what creating the marker looks like in code. Start with the basic Mapstraction map you created and add these lines to the create_map() function:

marker = new mxn.Marker(new mxn.LatLonPoint(37.7740486,-122.4101883)); 
// marker options will go here 
mapstraction.addMarker(marker);

The first line creates a marker object, passing latitude/longitude coordinates for the No Starch Press offices in San Francisco. By drawing attention to the graphical marker, we are essentially marking that spot as important.

The second line is a placeholder for any marker options we want to add later. (Any JavaScript line that begins with two slashes is a comment, and the browser ignores them.) The marker options are where we tell Mapstraction which icon to use or add a message to be displayed when the marker is clicked.

Finally, the third line adds the marker to the map. Once this happens, no additional options can be added. The reason is that the marker object is used only by Mapstraction. Once the marker is added to the map, however, Mapstraction makes the appropriate calls to the mapping provider. Mapstraction plots the marker based on all options set beforehand. In this case, we don’t have options to add, but we’ll add to this map in future projects.

If you’re using Google as your mapping provider, your new map will look like the picture below. The default Google icon sits in the center of the map. Although the marker is clickable, this marker is very simple and nothing actually happens if you click it.

Create a custom icon marker

The quickest way to make a map feel like your own is to change the default icon used for markers. Mapstraction has simple marker options that make the technical process of using custom icons a cinch. The more laborious part may be creating the icon file itself. To avoid this, you can find icons others have made online for free. I list several resources on my website.

Still want to create your own? Read on.

To create your own marker icon, you just need to have a graphics program that can save a transparent .png file. The icon can be whatever size you want, but keeping each dimension between 20 and 50 pixels is probably best. If the icon is too small, clicking it becomes difficult; too big, and the icon obscures the location you’re attempting to call out. If you’re using Google as your mapping provider, you also want to create an image to use as your marker’s shadow. This step isn’t necessary if your marker is a similar shape to the Google default or if you’re using another provider.

Not much of an image magician? Use the free online Shadowmaker service to create a shadow.

Add your icon to the map

Now that you have an icon, the easy part is adding it to the marker options. All it takes is setting a few values to tell Mapstraction where the icon image files resides. Your best bet is to keep custom marker icons in a special directory on your server. If you’re testing locally, you can use local copies, accessed by their location relative to the page containing the map. For simplicity, I have the HTML file and the icon files in the same directory in this example. In reality, you might prefer to be more organized.

I decided to use a teensy No Starch Press logo for my custom icon. It’s 27 pixels wide by 31 pixels high. Like I said, the icon is teensy. Then, I used the Shadowmaker service to create a file that is 43×31 including the marker’s shadow.

Finally, it’s time to code. Add these lines as marker options. These lines are inserted after a marker has been created but before the marker has been added to the map:

marker.setIcon('nostarch-logo.png', [27,31]); 
marker.setShadowIcon('nostarch-shadow.png', [43,31]);

The only parameter you need to include is the path to the image for both the icon and the shadow. Notice that the dimensions of each graphic get passed as an inline array. This parameter is optional but recommended. If you leave it out, some providers will assume the dimensions of the default marker, which could mean a poorly scaled graphic.

The results of the custom marker code are shown below.

The No Starch Press office is marked by the company’s logo, a little iron icon. Notice the shadow, as well, which makes the graphic pop out from the map.

Omit the shadow icon at your own risk. Some mapping providers will assume the default shadow, which might look silly with your icon. Not every mapping provider uses shadows, but planning for one is good. If you really don’t want a shadow, consider using a completely transparent graphic. I show an example of shadowless icons in the weather map example in chapter 10 of my book.

See also:

Where 2.0: Geomena Launches API to Feed its Open Location Database

Michael Calore — Wed, 31 Mar 2010 23:02:02 +0000

SAN JOSE, California — A new web service called Geomena is trying to build a geolocation database practically from scratch, and it’s taking a page from Wikipedia’s playbook to do it.

Geomena is an open wi-fi geo database — using a method similar to services from Skyhook and Google, any app plugged in to Geomena can use nearby wi-fi access points to determine your location.

The database is tiny right now. It has around 3,400 geo-tagged access points in the system, most of them around the project’s home base in Portland, Oregon. So, to grow the database as quickly as possible, the Geomena team has launched a new API that lets developers build apps that can enter new wi-fi access point locations.

So, if you’re making a location-based game, a location-sharing Firefox plug-in, or a web-app that relies on geodata, you can rig it up to write new wi-fi location points directly to Geomena’s database, helping it grow through good, old-fashioned crowdsourcing.

The emergence of location as an application platform has led to a bevy of new web services, each of them eager to provide developers with geodata to fuel the current flood of mobile and web-based apps. Most of the buzz at the all-things-location Where 2.0 conference, taking place here this week, has centered around SimpleGeo, a new web data store that just launched its “iTunes for geodata” — a pay-as-you-go solution for developers building location-based apps.

But iTunes isn’t the only model for building a powerful geodata store. Geomena is instead following the Wikipedia model. They’ve made their database entirely open and editable, and they’re licensing all of the data under the Creative Commons.

You could always just add points manually, but with Geomena’s new API, the system is open for developers to create new ways for users to plug in new access points.

Geomena’s Adam DuVander, who announced the API here at Where 2.0, imagines that most developers will build wi-fi access point discovery features into mobile clients or browser plug-ins. He also envisions various games that reward the user for feeding the system with points or badges. [Full disclosure, DuVander used to write for Webmonkey and is still an occasional freelance contributor to Wired.com].

We could even see war-driving apps being built that would log location data into Geomena for each sniffable wi-fi access point it comes across.

You may be wondering what the need is for wi-fi location sensing when so many smartphones and laptops have GPS built in. But consider how unreliable GPS becomes once you move indoors — it isn’t always an option in an office building, inside an apartment, or in highly congested urban areas. Also, for the near future at least, there will be a class of mobile and tablet devices that don’t have GPS. They will continue to rely solely on wi-fi for location sensing. The cheapest iPad (which will probably be the biggest seller) falls into this category.

If you live in Portland, you can make Geomena your location provider in Firefox right now. If you live outside of Portland, that won’t do you much good. So build an app that adds your local wi-fi nodes to the system. Water the garden, see what grows.

See Also:

Where 2.0: SimpleGeo to Launch ‘iTunes for Geodata’

Michael Calore — Wed, 31 Mar 2010 18:36:38 +0000

SAN JOSE, California – If you’re building an app that incorporates location — whether it’s a game, a local search service, or even a Twitter client — you’re going to have to go somewhere to get your data.

As we noted Tuesday, location is now an application platform, and there’s a whole crop of location data stores opening up to serve the emerging market of applications.

SimpleGeo is the latest such company to join the scrum. The web startup is announcing the debut of its geodata service here at Where 2.0 on Wednesday afternoon, but Jenna Wortham of The New York Times leaked the news a little early.

From the NYT Bits blog:

The company has been working to create what he describes as “iTunes for geodata.” The idea is simple: Create a wide sampling of geographic datasets and technologies that developers can access free or, for heavier users, at a range of prices. [...]

The company offers two tools. The first is the SimpleGeo Marketplace, which gives developers access to different location datasets and technologies for a monthly fee. The second is called the SimpleGeo Storage Engine and allows developers to perform location queries on a pay-as-you-go basis.

To gather its data, SimpleGeo began consuming datastreams from Twitter, Gowalla, Foursquare, Brightkite, Flickr and other location-sharing web services.

The pay-as-you-go model will work well for SimpleGeo, which allows the first million API calls for free, according to TechCrunch. Prices then start at $300 for the next level and go up from there. The company claims to have over 4,000 partnered developers using its service.

See Also:

Where 2.0: Fly a Drone Helicopter and Fight Killer Robots With Your iPhone

Michael Calore — Wed, 31 Mar 2010 17:12:28 +0000

SAN JOSE, California — First unveiled at CES in January of this year, the Parrot AR.Drone is a flying wireless toy that’s the center of a new augmented reality game. It streams video and sends location information as it hovers and zips around, and you can control it with your iPhone or iPod Touch.

As you control it, you see the drone’s POV video stream on your phone’s screen. Tipping the phone in different ways makes the drone turn and fly around, as the software senses the iPhone’s accelerometer.

As if a remote-controlled helicopter isn’t cool enough: The Parrot drone’s control screen has cross hairs, and you can “shoot” at things you see on the screen. The drone detects tags that people have applied to inanimate objects, and as objects are tagged, they can be replaced on-screen by virtual objects. So, as you fly around, you can shoot at virtual killer robots that are layered over the real-world background video. You can also put two drones into battle mode and shoot at each other.

Martin Lefebure of Parrot, the company that makes the device, demonstrated the latest version of the drone on stage at the Where 2.0 conference here Wednesday. The thing flew around the room, and everyone in the audience was able to look up onto the big screen on stage, where they could see themselves waving at the drone’s video camera. Lefebure then did battle with some insect-like evil robots that were holding us hostage in the conference ballroom. Unfortunately, he got his ass handed to him.

Parrot first showed off its iPhone-controlled car — the first concept that eventually evolved into the Parrot — at the 2009 edition of Where 2.0.

The iPhone and the drone talk to each other over a standard wi-fi connection. It has a range of about 150 feet (it’s limited by the range of your wi-fi) and the battery lasts about 15 minutes.

Location Isn’t Just a Feature Anymore, It’s a Platform

Michael Calore — Tue, 30 Mar 2010 23:01:12 +0000

The O'Reilly Where 2.0 Conference runs through Thursday in San Jose, California.

Just when you thought the swell of popularity around location-based services has hit the high water mark, the tide keeps rising.

All of the major web search engines are location-aware. Twitter has its own geocoder and Facebook is including location data in status updates. The big photo-sharing services like Flickr and Picasa support geotagging. Social location apps from Foursquare and Yelp are all the rage, and augmented reality is being touted as the next big thing. The emerging HTML5 specification has its own geolocation controls that let webapps find a user’s location through the browser.

We’ve reached the point where the addition of location data inside an application isn’t a special “bells-and-whistles” add-on, an experimental feature or a layer that’s only useful to some users.

It’s a standard feature now, and it’s crept into every product we care about.

“Location is something that people are just going to expect from now on,” says Brady Forrest, program chair for the O’Reilly Where 2.0 conference, the three-day event about all things location-based taking place in San Jose, California this week.

The location revolution was fueled by the proliferation of geo-enabled devices, Forrest says. Since most of us are carrying GPS devices in our pockets (every iPhone and Android phone has one, and most notebooks, too), it’s created a whole new application platform on which companies from different sectors — search, mapping, gaming, social networking, location-sharing — can compete.

“The platform is here,” he says. “Now, people are finding new ways to exploit it.”

Behind the scenes, there’s also a vibrant new market around web services for location-based apps. Cloud-based services like SimpleGEO, GeoAPI, Yahoo GeoPlanet and Skyhook are powering the most popular of the location-based apps we’re running on our phones and on our desktops.

Several of these companies will be at Where 2.0, along with Google, Microsoft, Foursquare, Gowalla, Yelp, Twitter — a who’s who of location-aware players. The three-day schedule — which features break-out sessions for developers on Tuesday, then two days of plenary on Wednesday and Thursday — includes discussions, panels and presentations from all the big names.

One of the most exciting things we’ve seen emerge from Where 2.0 over the years is the increasing importance of open source mapping systems like OpenStreetMap and OpenGEO. Such factors as the recent earthquakes in Haiti and Chile, the emerging mobile markets in developing countries in Africa and Asia, and the recent data transparency movement within our own U.S. government has proven the need for such crowdsourced and community-driven geodata services. They give anyone, anywhere the ability to get accurate place data and build it into applications quickly and easily, either for free or at a low cost.

Webmonkey will be at Where 2.0 Wednesday, March 31 and Thursday, April 1. Look for our coverage right here on the blog under the Location category.

See Also:

Facebook Finds its Place in the Location-Sharing Landscape

Michael Calore — Wed, 10 Mar 2010 00:33:14 +0000

The biggest social network on the web — that’s Facebook, by the way — is getting ready to unveil a location sharing service of its own, according to a report Tuesday.

Citing unnamed sources, The New York Times’ Bits blog says there will be two components, “a service offered directly by Facebook that will allow users to share their location information with friends,” and a set of APIs other location-sharing services can employ to allow Facebookers to update their location info using outside services.

NYT‘s Nick Bilton says Facebook will shed light on the new service at the company’s upcoming f8 developer conference in April.

Facebook has certainly taken its sweet time getting in on the location-sharing game — services like Foursquare, Gowalla, Google Latitude and Yahoo Fire Eagle have been blowing up over the last year. But the whole idea of “check-ins” raise new privacy concerns for many social network users. Some view it as over-sharing, others have concerns about invasion of privacy or cyberstalking — which is why all of the most popular location-sharing apps have extensive privacy controls built in to their opt-in services.

Earlier this week, reports surfaced that Google is experimenting with rolling location-based features into Buzz activity streams, and that the company is even working on a new location-based ad format.

Photo: Mr Ush/Flickr/CC

Google Gets a New Geocoder

Michael Calore — Wed, 10 Mar 2010 00:00:47 +0000

Google has announced a new geocoding web service app authors can use to better plot locations on a map.

The new Google Geocoding Web Service includes some enhanced capabilities that not only make it possible for app developers to provide more accurate and granular locations in their apps, but it also lets them increase the performance of their apps through precaching.

First off, the new service employs the Google Maps JavaScript API version 3, which has a handful of improvements over the previous versions. Users will get more well-formed and easier to parse data from each request. The service can return full names as well as local-language abbreviations for countries, states and territories. Users also have the ability to apply multiple tags to each address component.

Second, the new service lets apps precache data. From the announcement on the Geo Developers blog:

The Geocoding Web Service is intended to enable precaching of geocoder results that you know your application will need in future. For example, if your application displays property listings, you can geocode the address of each property, cache the results on your server, and serve these locations to your API application. This ensures that your application does not need to geocode the address of a property every time it is viewed by a user. However we do ask that you regularly refresh your cache of geocoder results.

It’s important to note that the new service must be used in conjunction with a Google Map, generated either by the Google Maps API or the Google Earth API.

Mobile Browsers Deserve Location Data

Adam Duvander — Tue, 25 Nov 2008 23:06:23 +0000

How come my desktop browser can get my location, but mobile Safari on my iPhone acts as if it doesn’t know where I am? I hate to seem impatient, but there is a proposed geolocation standard. Mobile browsers should adopt it swiftly.

The mobile web is here. The iPhone and Android are going to duke it out, and the end result will be more users. Facebook’s mobile visitors have tripled in the last year. That’s a lot of mobile browsing.

If the location-aware services are going to be as disruptive as everyone has said, these devices need to get better at sharing the information available within them. Yes, Android and iPhone both have apps. But we shouldn’t need to wrap our web projects in an app just to access the coordinates.

The Geode plugin for Firefox and its presence as a full feature in the browser’s most recent beta have proven it’s reasonable to include it even on a non-mobile machine. Despite the flakiness of WiFi-based geolocation, innovative sites have incorporated the technology. You can shout your whereabouts or tie files to a location all with the help of browser-based geolocation. Of course, we have a Geode/Gears geolocation tutorial so you can incorporate it, too.

But we really want it incorporated in mobile devices, so we’d be able to see some real innovation. Location-based services are at the horse and buggy stage right now. Let’s give it an engine.

See also:

Side Project to Startup: Shizzow Q&A

Adam Duvander — Tue, 18 Nov 2008 18:30:27 +0000

Shizzow is a new location-based social service, most similar to BrightKite. The bootstrapped startup is also a side project. The four team members have full-time jobs outside of Shizzow.

Webmonkey got together with Shizzow CEO-by-night Ryan Snyder. Read on to find out why he won’t make an iPhone app, Shizzow’s relation to Google Calendar, and that the original name rhymed with “kazoo.”

Webmonkey: How does Shizzow compare to other location-based social networking sites like BrightKite, or a platform like FireEagle?

Ryan Snyder: We think of Shizzow as much more of a social service than a location-based service. The primary action on Shizzow is to “shout out” your location, but to us declaring your location means nothing unless doing so enables you to get together with people for a face-to-face conversation. While other services have added photo sharing or restaurant reviews to their service, we’re maintaining a philosophy of simplicity – if a certain feature doesn’t help you meet new people or get together with friends, we won’t implement that feature.

Webmonkey: Many successful web applications start with developers scratching their own itch. How was Shizzow born?

Snyder: Shizzow is definitely one of those projects that came out of us developers scratching our own itch. In September 2007, a number of developers here in Portland were using a shared Google calendar to coordinate meetups for coding sessions, and we found it too cumbersome to notify each of the group members when we’d arrived at that place. Mark Wallaert approached me and said, “So… Ryan, I’ve got this idea…”, then sketched out the Shizzow concept on my markerboard. When he told me how it would solve our communication problems, I was sold.

Webmonkey: Why are you opening in only a few cities?

Snyder: One of the difficulties of unveiling a new site or service is that of building community. Rather than inviting random people from all over the world, we felt it would be better to invite people to use Shizzow city-by-city so that when we roll out to your city, all of your friends will hopefully be Shizzow users within the first day or two instead of straggling in over the coming weeks or months.

Webmonkey: How have the four of you balanced this large side project with day jobs?

Snyder: Whew, this has not been easy! I’d probably call it “burnout prevention” before I’d call it anything resembling balance! Each of us has our own methods of meeting Shizzow’s needs on top of our day jobs. I personally dedicate the first 2-3 hours of my day to Shizzow before heading into cubicleville for my 9 to 5′er, as well as dedicating one or both weekend days to whatever tasks may be at hand. But the real reason we’ve been able to persevere over the last year has been the patience and understanding of our friends and loved ones. We simply could not have done this without their support.

Webmonkey: BrightKite got a lot of attention for its iPhone app. When can I expect to see a beta version of Shizzow’s?

Snyder: Since we’re a small team, we’re trying to remain as focused as possible on Shizzow’s core functionality. We feel that developing platform-specific applications will actually scatter our attention by having to support multiple UIs and platforms. We’re currently working on an API to allow other developers to build applications for Shizzow. Besides, there are some rockstar mobile developers that will probably build something cooler than we’d imagined possible using our API.

Webmonkey: My projects always have a list of alternate names. Can you share anything Shizzow was almost called?

Snyder: All of our original names for the project were either taken or they were just lame! Our first interface for Shizzow actually had a spelling variation, where we ended Shizzow with “ou” instead of “ow”. People kept calling it “Shizzoo” so we quickly realized that we needed to grab the “ow” domain name before that name stuck!

[Photo by Aaron Hockley]