2nd Zooniverse Project Workshop
Today we have a guest post from Katy Maloney who attended the 2nd Zooniverse Project Workshop hosted in Chicago at the Adler Planetarium (US Zooniverse HQ) earlier this week. Katy is a full time sociology student at University of Quebec in Montreal, with a certificate in religious studies. Her main research interests are STS, citizen science communities, man/nature relationships and ontologies. She’s also a part-time Planet Hunter(ess).
When I was invited to the Zooniverse workshop, I initially thought I would be kind of like a fish out of citizen waters: I was basically expecting a tech-fest, and a few scientists looking at me like the odd one out, an undergrad student in a sea of professors. Well I was right about being pretty much the only undergrad, and that’s about it! As previously mentioned in the Planet Hunters forum thread, the main focus of this workshop was community. That means you guys. Or us guys. First of all I felt very welcome, grabbing dinner with Jules the night before and then a few beers with Chris Lintott, Lucy Fortson and a few other zooniverse peeps. It was great to have Jules as a guide of sorts, introducing my somewhat shy self to people she recognized in the dark meeting room in the basement of the Adler Planetarium. It was so dark in there, in fact, that I mistakenly traded my usual morning coffee for a decaf without even realizing it! Lucky for us all, the auditorium was better lit, and everyone introduced themselves prior to the day’s events, so that took care of the “room full of strangers” issue! And as the first few morning conferences went by, I was more and more convinced I was not only welcome, but I was exactly where I should be and would find opportunities to participate in the sessions later.
A first welcome speech by the Adler’s new leading lady, Michelle B. Larson, who is a trained astrophysicist with a passion for science education and public outreach. The Adler’s support of citizen science will no doubt continue to flourish under her leadership. Then Chris Lintott, PI and omnipotent leader of the Zooniverse, gave a talk about the history and evolution of the Zoo for us all to enjoy. What was originally a couple of loosely coupled projects turned into double the projects fairly quickly: if it were a reality TV show, it would probably be named “14 Zoos and counting”, as new projects are already underway and even as we speak, there are enough proposals from the science community to double the projects we currently have!
As of now, we have over 800,000 volunteers signed up to analyze and classify various scientific data. As every other project goes live, other pre-existing projects also benefit from the growing “fanbase” of the Zooniverse. People come online to these projects to do authentic science, and emphasis needs to be put on the word authentic. The Zooniverse provides an environment where scientists can focus on the science, all of the engineering, designing and promoting platforms already made available for them to push forward their science objectives and present them to the citizen science world. Projects are viewed as a learning experience for citizen scientists, but also for citizen science: that’s why each project is different, new, and Arfon Smith, director of citizen science at the Zooniverse, will elaborate on that later in his talk. The “one visit rule” was probably derived from knowledge of what did and didn’t work in previous projects: people need to be able to start doing science on the first day, after minimal training time. This was stressed many times by Meg on the forums, and is later explained further in Karen Carney’s talk with statistics gathered from Zooniverse projects: a great amount of users are one-timers, and if we are to benefit from their visits, we need to have this affair leveled to their needs. All in all, complex doesn’t have to mean complicated, and importance was being focused on feedback: what is the “bacon”, the motivation that keeps the machine running? Is the correlation between citizen science and science literacy one of cause and effect, or were zooites simply already scientifically inclined? How can we foster serendipitous discoveries, or in the very least, provide the right kind of communication channels so that we intercept them as they happen? Simulated data, ZooTools, beta testing, communication and promotion tools, all of that good stuff was on the table to discuss later in the un-conference sessions.
Arfon was the second Zooniverse leader to take the stage and present a morning talk, titled “Capabilities of the Zooniverse: What we can do and what we’ll be able to do”. In 4 bullet points, he described the efficiency method developped by Zooniverse over the years: 1) Domain model, 2) Technologies/Tools, 3) Data in/out and 4) Intelligent systems. Seems a bit out there and unclear to the layman, so let me elaborate a bit. The first point has to do with a vocabulary the developers, science teams and other people involved in the Zooniverse have in common. There are subjects (things that people analyze ), tasks (things that people do through the classification interface), and users (volunteers, people). The second point has to do with a certain open source ethos: picking hard and different projects, testing the limits of what we can compute in a simple and sleek interface, the scale at which we can expand (in terms of traffic and support of huge inflow of data on the servers), in accordance to a strong science commitment (producing the best science we “citizenly” can). He had some info thrown in there on the subject allocator/classification API, that I will let the upcoming videos explain, as I’m not too sure my explanation would be any more enlightening, this is a bit over my head in terms of computing knowledge. It has to do with a core application called Ouroboros.. The third point is a bit fuzzy in my memory, but was about actual physical data I believe, pictures were thrown in of people with boxes and boxes of hard-drives (one of the many joys of metadata analysis I guess!). Data reduction was also included into this, while the fourth point, intelligent systems, described how data gathered on user behavior could inform and allow for fine tuning, eventually leading to a strength-oriented classification which would act as a potential negation of the equalization effect of the interface, which is tuned to accommodate first time users and “metausers” alike.
Speaking of the devil, who exactly IS the Zooite?! According to Karen Carney and Stuart Lynn’s accounts, his name is David Smith, he’s a white male, 40-41 years old, educated, he has a technical job and is from the US. Oh, and he wears glasses?! (I hope Stuart’s talk will be in the broadcast version, it was hilarious! Apparently a lot of zooites love their mothers.. or yours, more accurately!) Karen’s talk was jam packed with statistics, that I can’t all mention here, for lack of being able to write them down quickly enough! Most notably, their research have identified 3 types of users: initial users, those who visit only once (roughly 39%), sustained users, those who come back many times (59%), and metausers (2%). I was almost ready to make up slogans like “We are the 2%!” when I realized I was probably more likely in the 59% category myself. Turns out I’m a pretty unremarquable zooite according to those statistics! And that sits well enough with me – defying classification! 😉
They also identified 3 “whys”, 3 bacon strips, 3 main reasons that zooites express as their motivation for joining a zooniverse project: social engagement, content engagement, and scientific engagement. Remember when we were asking if committed users were “converted” to a scientific interest, or if they just happened to be interested already? Well the demographics are consistent with science-attentives, meaning people who were already interested in science, be it through reading science mags, science news online or through a job that was technical and already sorta science-related. This relates to an issue that was brought forward: the confidence barrier. People have to believe they CAN do science. In fact, most people who think they’re doing terrible are doing fine, and some people can inversely get pretty cocky and think they got it down, while being completely off-track (although luckily perhaps, they are seemingly in lesser proportion). It’s the team’s prerogative to make sure that while more science-attentive people feel like they’re doing legit, real, authentic science, people who have no prior science background also feel like this is not above their heads. So the data presented was essentially quantitative (statistics, numbers based on surveys), but qualitative research (the “whys”) is also being conducted, by the Adler team, including Mr. Ryan Cook, which is a socio-psychologist (or should we say psycho-sociologist?! Sounds a lot funnier.. hehe) working on citizen science learning, and by another team based at Syracuse University, working on zooite motivation (again with the bacon!).
So what IS an un-conference exactly? A colleague of mine in sociology was joking around a few weeks ago telling me how un-conferences were essentially conferences where everyone could talk over each other and argue to no point, a sure-perfect way of wasting time, rubbing shoulders and shaking hands at high cost. While I agree that in some contexts it might be so, I found that in the context of this event, a citizen science assembly with many different people from different projects getting together to discuss things they have in common, it was more like a workshop/discussion group exercise, allowing for diversification and exchange on perhaps more circumscribed subject matter. People proposed subjects to discuss, over lunch, and then voted on their favorites in subjects submitted. The most popular subjects each got a session in one of the two afternoon time slots, which could each accommodate three to four different matter subjects. I mentioned that I met up with Jules, moderator for a few Zooniverse projects including Moon Zoo, for dinner on Sunday night. We discussed an un-conference session there, that she aptly titled “How to keep 800,000 research assistants happy”. Why research assistants rather than volunteers, user or zooites, you might wonder? It stemmed from a scientists working on a zooniverse project who once told her, after seeing his project go live, how great it was to suddenly have 10,000 research assistants working on his data. I thought it was also bridging the gap between the science teams and the volunteer base, in an effort to push forward a collaborative mindset. In this session, we discussed this “bacon” that Chris was talking about, from a volunteer’s point of view, and issues that were brought up in the discussion boards prior to the event.
Janet’s recollection of the session is pretty spot on:
1) Don’t Waste People’s Time. Make a good case for the science. Make it bug-free or get quick fixes. Educate the volunteers.
2) Volunteers are Collaborators. Include volunteers on beta tests and development. Be honest about what is happening, and share data and results. Communicate, encourage and give recognition when it it due.
3) Only present tasks machines cannot perform. If data the data collected allows a good machine algorithm to be written, inform the members of their success before closing the project down as “finished”.
We’re not difficult, basically: what zooites really want is to feel that their work is worth something. After all the hours spent classifying, they want to be sure that their time isn’t wasted, and that they aren’t subordinates to a despotic science team, ready to shut the shop down unceremoniously as soon as they have their results. The experience of the Supernovae project was put forward, even though we know its quick demise was due to our effectiveness (yay?!) and to an exterior person going “hey, your data allowed us to build an algorithm that’s doing just as good as you are, so thanks, I think we’re done here!”. That machine-learning aspect was discussed in another un-conference session, which Jules and Janet both attended and will probably provide more details for in their respective blogs/forum posts. While “metausers” are only but a small fraction of all zooites, they represent the core of each project, they’re the “expert citizens” that can better help keeping the community happy and informed. Furthermore, according to Jules’ numbers, which were agreed upon by zooniverse peeps present, 10% of the users do 90% of the work. Should I take my proverbial “We are the 2%” sign back out of retirement?! Yes and no.
The general conclusion I’d like to draw on is that the Zooniverse is well aware of and interested in you metausers and other sustained users. Even if the interface is technically made for the casual visitor, you are not forgotten, and fun things are in the making (like ZooTools, mentioned on the forums). I attended a session on the new Talk interface (which you can browse on the Planet Four site, as well as all newer Zoo projects), and all that attended were really excited about this version. Talks of a “Talk Tutorial” were had, to allow people who (like me, ahem…) aren’t necessarily familiar with hashtags and ways to optimize your use of Talk. Good news is, Talk 2.0 will be coming to Planet Hunters, sooner than later! And we can expect annoying bugs (like the image resizing and “sticky posts” issues) to be resolved also in a timely manner!
I realize that this blog is getting quite long (I did warn Meg that I was a blogging machine.. haha!), so I’ll let you guys go read Janet’s resumés of the other un-conference sessions of the day over on the Old Weather forum (a copy is also in the Zooniverse Workshop thread in the PH Talk Chat section), and will write another blog for Day 2 of the 2013 Zooniverse workshop.
It will broadly describe my experience as a zooite at the conference, as well as summaries of the second day morning lectures, including several case studies, followed by an account of a Zooniverse advisory board meeting, which I attended with Janet and Jules over lunch. I then rushed over to an un-conference session on Translating Zooniverse Projects (being a frenchie, this just spoke to me), thus missing a session hosted by Ryan Cook on measuring science literacy, which I would’ve otherwise loved to attend, and then to a second un-conference about education and its inherent awesomeness (sic!), which I got to summarize (nervously) for attendees at the end of the afternoon.