Travel and mobile applications ought to go together like peanut butter and jelly, but the cost of international data consumption generally stands in the way of using your phone to document the places you’re going and the fun you’re having. Poor connectivity can be a real drag too.
New technologies are being implemented by app makers and platforms to leverage Wi-Fi and selective data transmission so that mobile users can have more fun on the go. We recently took a look at how three companies are learning to make mobile apps smarter about the data they require. Some people say that developer education on the topic of limiting demands on the network is imperative for the future of the app ecosystem; these three approaches could provide some useful inspiration.
HipGeo is a new location based social network lead by tech industry veterans, but with an emphasis on cutting edge features and usability. Launched last fall, the app offers persistent location tracking and automatic creation of a travel blog chronicling wherever you go. Unlike some of the most popular mobile apps, like Facebook and Foursquare, Hipgeo can function with very little access to data.
“From the beginning we knew that data connectivity would be spotty and uploads flaky,” HipGeo’s VP of engineering and ex-Yahoo exec Jeff Kunzelman said, “so we designed HipGeo from ground zero knowing we needed to cache and background upload our GPS data.”
“We had the need to stream large data sets to our
servers to maintain the fine-grained tracking needed for HipGeo’s ‘Places and Routes’ mode,” Kunzelman explained. “To reduce the required bandwidth and increase reliability, our app gathers chunks of data then sends it to the server in sensibly sized blocks. If the uploads fail or there is no connectivity our app backs off transferring the data and tries again later.“As HipGeo evolved, our GPS algorithms got much, much better at reducing the amount of data it actually needed to send. As a result we found we needed the cache much less, but many of our users were using HipGeo on an iPad2 without a data plan or using it without cell and data connectivity when traveling abroad. Some wanted to go many days at a time between Wi-Fi connections. By reducing our datasets and optimizing our caching we found we could allow the users to do just that, travel for a few days, with full ‘Places and Routes’ recording, then upload the data when they find a place with good connectivity.
“Our cache manager is pretty clever about how and when to upload GPS data and even larger image data. Also our server algorithms can consume data completely out of order and handle partially received data blocks … our server backend and website is a very significant and integral part of our platform as well. The backend does some very heavy lifting analysis of the data delivered by the app to determine what and how things appear in the user’s blog.”
Efficient geotagging of media assets created by users and reliable, low-cost location tracking make a number of things possible, says Dr. Phil Hendrix, founder of analyst firm Immr.
“I think the big opportunity is to use the data to personalize users experiences based on footstream X time profiles – e.g., ‘looks like you are in an area where you often stop for lunch,’ or ‘looks like you are visiting Home Depot for the first time;’ or (very interesting) ‘looks like you are in an unfamiliar area …, would you like to see what locals recommend?’ I would be surprised if this isn’t on [HipGeo’s] roadmap. Or, this ‘location awareness’ could all be in the background to help developers (and brands, retailers, content providers, etc.) personalize based on the person’s familiarity with the location, store, etc.
“[This could move] HipGeo in the direction of personal data vaults, which have their own issues, but I think in the future this concept will take hold. It has to … location data are too useful to let it go out the exhaust; someone has to figure out out to reassure end users that their information is secure; handle privacy, granularity, etc.
“Advertising and offers are rapidly moving toward real-time personalization. SoLoMo [social/location/mobile] digital signals, especially location, interests, intent, and context, will play an increasingly important role, boosting relevance for consumers and yield for advertisers and merchants.”
HipGeo competitor Geoloqi says it offers similar functionality in its location-tracking app and in the software development platform it makes available to other app makers.
“I’ve used the Geoloqi Android app to collect GPS data with no data connection while in Sweden,” says co-founder Aaron Parecki. “Our app does the work of batching the data and sending it to the server when there is connectivity again. We’ve known that offline use is important for quite some time, which is exactly why we built this functionality into our Geoloqi SDK. If you want to develop apps like this, our SDK handles offline tracking and sending later for you.”
HipGeo launched an SDK of its own in November.
Ted Morgan, CEO of location services provider Skyhook Wireless, says his company’s on-device platform serves up location data to mobile apps using two different methods.
The first method is local caching, but with consumption instead of publishing in mind. “When you request a location in a new place (you just landed in New York), we actually download ’tiles’ of reference data for the area around you – much like map tiles in Google Maps,” Morgan explains.
“These tiles are squares of Wi-Fi and cellular reference data for all the access points and cell towers within say a 100 meter by 100 meter area. If you stay within that area we never go back to the server, just use the data on the device to calculate your location. As you move out of the tile into an adjacent tile, we will then request new tiles from the server. All for efficiency.”
The second method Skyhook uses is batch offline location requests. “Some apps and devices are mostly offline and only periodically have a network connection,” says Morgan.
“Cameras are a great example, also RFID tracking tags. The user doesn’t need to see the geotag for a picture they take immediately, they just want it recorded for the future viewing if their picture gallery. In those cases when someone takes a picture, we take a snapshot of the wireless environment around them at that exact time, all the Wi-Fi signals, cell towers, GPS signals and we store them with the image. When the user connects to their laptop or an access point to upload their photos at that point we take the snapshot data, send it to our servers and convert it into the latitude/longitude of where that picture was taken. The user never even knows it is going on but now the photo is properly geotagged.”
Whether it’s incorporating these kinds of methods into original software development, leveraging a third-party API or using location data provided by on-device software like Skyhook – you can expect many mobile apps to begin incorporating technologies that will allow them to ask much less of the wireless network and thus be used much more by consumers.
