Citizen Science Month, held every April, is a month-long celebration of citizen science, where volunteers from all walks of life get involved in research by collecting data, analyzing results and helping solve some of the biggest problems in science.
Citizen science reportedly dates back to ancient China, where locusts, who were migrating, wreaked destruction on harvests and residents helped to track outbreaks for over 2000 years. The phrase citizen science took a lot longer to catch on and was not coined until the middle of the 1990’s by sociologist Alan Irwin who defined it as, “a form of science developed and enacted by the citizens themselves”.
Since then public engagement in citizen science has grown to cover a whole range of activities from studying images of space to counting migrating birds, and by doing so the general public can help to contribute to our understanding of life, the universe and everything!
This week as an ode to Citizen Science Month we caught up with Dr.Coleman Krawczyk, expert researcher in data analysis for Zooniverse citizen science projects at the University of Portsmouth about what is possible in the field today and his hopes for the future of citizen science.
Who are you? What do you do?
My name is Dr Coleman Krawczyk and I am a postdoc at the Institute of Cosmology and Gravitation at the University of Portsmouth and the main focus of my research is working with Zooniverse Citizen Science Projects, specifically looking at the data processing of these projects.
Can you explain Zooniverse?
To give a brief overview for anyone who is not familiar with the Zooniverse, essentially research teams from all over the world can partake and build projects on the Zooniverse website to enlist the help of the general public in their scientific project.
When they do this they set each project up as a series of tasks to be completed, so that volunteers (members of the public) can participate by spending time completing these tasks (which could be anything from multiple choice questions, to drawing shapes on an image depending on the project).
Each piece of data needs to be checked by lots of volunteers (between 5-15) and then a consensus answer needs to be reached – which is where I come in building a tool box that the researchers can use to take these multiple answers, and turn them into a consensus. (a first step in that direction anyway!)
What’s the best thing about your job?
The thing I love about the job is the variety of projects that I work with, there is always something different, I am always learning something new.
I am an astronomer by training, but coming into this I have worked on a whole host of projects ranging from text transcription, in which volunteers were asked to transcribe handwritten letters from the civil war era in the USA – which would make these texts more useable and accessible for researchers of that time period.
To dealing with ecology projects, with my favourite being, penguin watch. For this project they were studying the life cycle of penguins in the Antarctic, to assess fluctuations of the population and how that changes species to species. To do this, they set up camera traps in the Antarctic, that take an image every few hours, at a couple of different times of year and the volunteers are tasked with a simple population count in which they view the images and are asked to ‘click’ on each penguin in the image. This data is then collected up, which is where I come in, to turn the clicks into a number of penguins. I was then invited to Antarctica to help collect the data, who could have thought that was possible with a Phd in Astrophysics!
What are your favourite citizen science projects?
From the space side of things the main project I am involved in and love is the project that spawned the Zooniverse in the first place called Galaxy Zoo. It started in July 2007, with a data set made up of several million galaxies imaged by the Sloan Digital Sky Survey (SDSS Telescope), it asked volunteers to help explore galaxies near and far, sampling a fraction of the roughly one hundred billion that are scattered throughout the observable Universe by assessing the images taken from the SDSS, and identifying particular shapes volunteers were able to help visually classify and identify galaxies and their positions etc.
Each one of the systems, containing billions of stars, has had a unique life, interacting with its surroundings and with other galaxies in many different ways; the aim of the Galaxy Zoo is to try and understand these processes, and to work out what galaxies can tell us about the past, present and future of the Universe as a whole. The volunteers were so important here because at the time this was not able to be done by a computer (although this is different today).
This project is still going strong, and I have been involved most recently updating the website code to migrate to the new Zooniverse system, which enables any researcher to join and start a project.
Why are citizen science projects important?
From a general perspective, it really helps to break down the barrier between the public and scientists – anyone who participates in a citizen science project, is doing real science, they are no longer observing, they are the scientist.
For example within Galaxy Zoo a number of astronomers on the project, said the only reason they studied Astronomy is because they were introduced to Galaxy Zoo when they were in highschool and it sparked that imagination, which meant they were able to see themselves as Astronomers for the first time.
I think this is a very powerful thing across many of the projects that those actually participating have a sense of “I did that” or “I helped with that”. To that end, in all our publications we always reference the volunteers and credit their achievements.
What’s the future for zooniverse / citizen science?
Some of the most exciting things happening right now is the combining of machine learning and citizen science. When the Zooniverse began, machine learning was not good enough, there were not the computer techniques to viably help, but now it can – and we can train the computers using the years worth of data sets that we have (for example with the Galaxy Zoo project).
And to some extent on other projects we are looking at more general ways of integrating machine learning with citizen scientists. Our main mantra of doing this is to ensure we don’t waste our volunteers time – so do we need to be collecting 15 classifications on each image, or can we get the same answer with 5? And we are looking at bringing in machine learning and seeing how the computer looks at the information / tackling the data from a more information theory perspective. All of this will give us an idea of how to better make use of the data we collect.
Ultimately the same thing that worked for Galaxy Zoo with the SSDS, will not work for the next telescope, to give you an idea LSST Telescope in Chile will be collecting the same amount of data that was previously collected in 10 years, in 3 nights. This far out-strips what volunteers could identify alone, so we have to utilise the tools available to us to ensure that we have better answers in the future and a team of volunteers and machine is best placed to do that – to ultimately get better results quicker!