The Social Web of Crowdsourcing

The Galaxy Zoo is a web site that contains millions of images of galaxies. These images where taken with a robotic telescope and they show pictures of galaxies in different form. Some are spiral, some are disk shape and some are rounded and some boxy. The Galaxy Zoo team has a problem: they need to classify this huge amount of pictures. Surly, it will take them years. But the team has found another way. They are asking normal web users to help out and identify the shapes in the pictures. This is an example of crowdsourcing.

When signing up, the participants look at the pictures and classify them. They identify the images according to the guidelines provide by the site. The results? According to the site, within 24 hours of launch, the site was receiving 70,000 classifications an hour, and more than 50 million classifications were received by the project during its first year, from almost 150,000 people.

Without the Internet this would be impossible. As the cost of coordination has dropped with the internet, this approach becomes possible.

Few thoughts are quickly raised about crowdsourcing. First, how can the site owners allow anybody – amateurs if you will, to classify galaxies, a task that can only be performed by the trained and professional eye of an astronomer. As it turns out, shapes of galaxies are not that hard to recognize. On August 2, 2007, Galaxy Zoo issued its first newsletter which explained that 80,000 volunteers had already classified more than 10 million images of galaxies (see Wikipedia entry on Galaxy Zoo). And it turns out that the amateurs are pretty good at it.

The second thought raised is usually about trust. What if someone with evil intent deliberately enters wrong classification and thus vandalizes the project. This is of course always a threat that is enough to kill any new effort. But as it turns out, this is not a huge problem. In fact, people are really just helping, volunteering their time for a good cause. While vandalism may be a nuisance to site like Wikipedia, the people of Galaxy Zoo can avoid the problem and decide to use only the images that have been processed by a specific number of people that all agree on the classification.

The third thought has the most surprising results. Why would people spend their time on identifying images of galaxies? Do people not have anything better to do, and this sounds boring anyways – and you don’t even get paid! To much surprise, people are actually willing to devote their time and effort to do this. They don’t have to. The are many strange forces that work here. Simply put: they do it for themselves, in their own self-interests. People usually enjoy voluntary work, it makes them feel good. And they can stop anytime. If they got paid it would be work and a boring one too.

With the Internet, sites like Galaxy Zoo can tap into a huge resource: people connected to the Internet, a pool of billion people. The numbers speak: 150.000 people devoted their time to classify 50 million images. The social web is changing the game. Changing how we  use the web and what is possible.

Galaxy Zoo is only one example of web sites that are bringing the work to people using the web. Wikipedia is the classic example. Hobby focused sites Flickr and YouTube are other examples. Google is also strong in the crowd sourcing field. Good example is the Google Map maker. With this site, users can add features to a map such as roads, labels, shops and even draw regions. Who would know a place better than the people who live there! I recently added couple of statues to a park near my house, Klambratún. Only the next day I got a comment from another user questioning my addition. So I sent some pictures to prove my point.

Even companies can use crowdsourcing as a business model. Classic example is InnoCentive that brings scientists and anybody interested together with companies to solve problems and come up with innovative ideas.

The traditional approach for research has been to form an internal group of people that will work on a give problem. And in many cases this have proved to be very successful. However there are some limitations to this approach. Companies can only enlist few people. They might be very intelligent but they will only have limited views and knowledge. If they could enlist everybody – or at least anybody, they would tap into the resources and knowledge of people that would be impossible to hire. InnoCentive claims to have enlisted more than 160,000 brigthest minds to work on solutions.

The social web is still in its early stages. We are only just beginning to understand the potential of human network connection and how people can interact.