Jetpac, a free iPad app that lets you peruse friends’ travel photos and build a bucket list of destinations to visit, raised $2.4 million in funding today from Khosla Ventures, Internet entrepreneur Jerry Yang, and Morado Venture Partners.
Now, the hipster-friendly travel photo app is encroaching on Tripit’s territory by helping users plan upcoming trips. With a new feature called “Atlas”, the app can tell you whether a friend has already traveled to a particular destination so you can ask them for handy travel tips and browse their photos.
“We built this search feature because users have been asking for it,” said co-founder, Julian Green in an interview with VentureBeat. Green, an avid traveler, told me he couldn’t understand why there wasn’t already an easy way to find out if your friends had already traveled to a destination before you fly.
The app has indexed over 750 million photos from Facebook. Green said it’s only a matter of time before Jetpac integrates with photo-sharing sites like Instagram and Flickr to build a database of over a billion travel images.
Since its launch in April, the app has consistently sat at the top of the “New Travel App” section of the iPad store and is known for its sleek, vintage-postcard-inspired design. Green said the founders share a vision to create “an old 1960’s carousel slideshow with all of your friends’ best travel photos.”
Given that the average Facebook user shares 200,000 photos, how does the app find the best ones?
Well, you’ve already done the work. Jetpac crawls the captions from every image you’ve shared on Facebook. If it’s a travel photo, the caption will usually reflect that as you’ll include location information like “Taj Mahal” or “night out in San Francisco.” So Jetpac’s algorithm manages to largely avoid pulling local information and images.
We also tend to “like” photos from our friends’ ordinary lives (a grainy picture from a Bob Dylan concert, for example), so the founders made the decision not to place any credence on whether an image has been highly-rated on Facebook.
The algorithm functions with over 90 percent accuracy in selecting high-quality travel photos. The founders developed it by rating thousands of photos by eye, and then used machine learning techniques to find common traits among good and bad photos. As an example, the algorithm favors captions containing Peru and Fiji, but usually dismisses photos containing the words “Mommy” and CEO”.
The tech is well-designed, but isn’t flawless and the app sometimes has trouble identifying places. Green admits that Jetpac often can’t tell the difference between Paris Hilton and people who’ve traveled to Paris, France.
The startup faces competition from Trippy and Gogobot, but the founders remain confident that theirs is the only app with a strong emphasis on images. If you’re bored during a commute, it’s addictive and fun to browse pictures of exotic places and fantasize about going there, whether you have any immediate travel plans or not. “Jetpac is a place to dream about your next destination,” said Green.
VB's research team is studying web-personalization... Chime in here, and we’ll share the results.