Zoom Out Me
Valentina Camacho
An experimental site that lets you explore recommendations outside the range of possible things you might like, by using the Hunch algorithm.
http://www.zoomout.me
We have been told that the Internet is a very broad world, where we can find anything but we end up finding the same things over and over again. This happens because search engines are recording our every click, and are making decisions about what we should see based on algorithms which are accurate most of the time.
ZoomOut.Me offers a search experience outside the user’s tastes by using the “Hunch Recommendation” algorithm. The point is to prove that our tastes vary and that we are unpredictable humans that can’t be measured by any algorithm.
Background
I did a lot of research about recommendation algorithm. I read Collective Intelligence by Toby Segaran and read the way Google created the page rank algorithm. I also speak with a lot of people and found that I wasn't the only one suffering "internet claustrophobia".
I did a previous work called "Terms I agree" in which people were asked to pay attention to what they were agreeing on the Internet. In the case of Zoom Out Me, I wanted to explore a different interaction, so I research about spirals.
Audience
Internet users that spent a lot of time in front of their computers.
People that look for recommendations all the time about book, music and movies.
User Scenario
The user sign up to Zoom Out Me by connecting through the Hunch algorithm and then results shown are matching the center of the spiral. The user then can start exploring for recommendations by moving a circle into the spiral. They will see things outside their range of taste and they can save the interesting things they found to review them later.
Implementation
The site was built in Heroku, using Node.js and Mongo DB.
The spiral was made in processing and the introduction and the coverflow are jquery plugins.
Conclusion
I learn how to build a dynamic website, specially the backend.
How to connect a processing sketch into a website.
How to work with APIs.
I discover that there are things that you can be surprise to find in your search and that it becomes funny to see what the algorithm is not normally showing me.
ZoomOut.Me offers a search experience outside the user’s tastes by using the “Hunch Recommendation” algorithm. The point is to prove that our tastes vary and that we are unpredictable humans that can’t be measured by any algorithm.
Background
I did a lot of research about recommendation algorithm. I read Collective Intelligence by Toby Segaran and read the way Google created the page rank algorithm. I also speak with a lot of people and found that I wasn't the only one suffering "internet claustrophobia".
I did a previous work called "Terms I agree" in which people were asked to pay attention to what they were agreeing on the Internet. In the case of Zoom Out Me, I wanted to explore a different interaction, so I research about spirals.
Audience
Internet users that spent a lot of time in front of their computers.
People that look for recommendations all the time about book, music and movies.
User Scenario
The user sign up to Zoom Out Me by connecting through the Hunch algorithm and then results shown are matching the center of the spiral. The user then can start exploring for recommendations by moving a circle into the spiral. They will see things outside their range of taste and they can save the interesting things they found to review them later.
Implementation
The site was built in Heroku, using Node.js and Mongo DB.
The spiral was made in processing and the introduction and the coverflow are jquery plugins.
Conclusion
I learn how to build a dynamic website, specially the backend.
How to connect a processing sketch into a website.
How to work with APIs.
I discover that there are things that you can be surprise to find in your search and that it becomes funny to see what the algorithm is not normally showing me.