Citizen Science Alliance

The Citizen Science Alliance's mission is to create online citizen science projects to involve the public in academic research. We believe that by doing this we can not only help everyone share in the excitement of discovery, but that such projects are a necessary response to the flood of data facing researchers in many fields. As data sets have ex... panded in size due to the rapid decline in cost of computing, detectors, bandwidth and storage, so traditional modes of research have struggled to cope. While machine learning and computing have been able to take up some of the slack, they are not always adequate replacement for human abilities. After all, our brains have evolved to be extremely good at pattern recognition and we should try and take advantage of this ability where possible. In the early years of this data flood, researchers were able to cope by recruiting more willing collaborators and students, but in many fields this proved to be only a stopgap. We need a much larger workforce than any academic department can provide. Luckily, the web provides a means of reaching a much larger audience, willing to devote their free time to projects which can add to our knowledge of the world and the Universe. We call this method of involving the public 'citizen science', and we believe working this way has several advantages, a few of which are listed below. The ability to cope with extremely large data sets - in its first six months Galaxy Zoo provided the same number of classifications as would a graduate student working round the clock for 3.5 years. Unlike work by a small number of experts, our ability to gather multiple independent interactions with the data provides quantitative estimates of error. This is an essential part of the 'wisdom of crowds', allowing us to understand the accuracy of the data we provide. Citizen science data sets naturally provide large and powerful training sets for machine learning approaches to classification problems. This is an essential part of our future; as data sets continue ...Read More