This is the title of an interesting paper by Carol Christian et al., in which they discuss the role of citizen science in modern astronomical research. The authors adopt the definition of citizen science as ” active involvement of a large group of individuals to achieve real research objectives with measurable outcomes.” Many contributors to citizen science projects are not in fact trained in scientific research, and contributions often center on research enabled by pattern recognition: the human eye is far better at this than a computer.
Perhaps the largest and best known collection of citizen science projects is the Zooniverse, the progeny of the Galaxy Zoo project, which realized 60 million visual classifications of galaxies measured by the Sloan Digital Sky Survey (SDSS). Galaxy Zoo has produced 21 papers in 3 years, and has received a total of 500 citations.
The authors argue that citizen science is a major component of large data set analysis in three ways:
- Scalability, brought about rapid penetration of high-speed internet connections in the past 10-15 years and the adoption of common astronomy data formats. Technologies such as cloud computing can deliver the needed resources on demand.
- Serendipity. The human eye is very good at picking out the unusual. One example is Galaxy Zoo Peas, first noted because of their green color in the SDSS survey, caused by very high efficiency star formation.
- Relation to Machine Learning. Citizen science projects can produce large training sets to improve machine learning algorithms.
Perhaps the most interesting part of the paper is the argument that citizen science can begin to play an important role in observatory data analysis pipelines. The figure below shows the data flow for a pipeline and where citizen science may be useful:
Data deposited in the observatory archive can be made available to citizen science projects such as those contained in Zooniverse. Results fed back to the archive can inform improvements in the scientific value of the archive.