“Inviting bots and citizen scientists into the National Maritime Museum” - Fiona Romeo talk at the Guardian

Martin Belam  by Martin Belam, 1 August 2011

Fiona Romeo, Head of Digital Media at the National Maritime Museum and Royal Observatory visited the Guardian to give us a lunchtime talk last week about “Inviting bots and citizen scientists into the museum”. It was fascinating - and delightfully geeky.

She spoke about the invention of the “astrotagging” convention for photographs, a sort of “geotagging”, but for the universe. This necessitated having rather more data than simple latitude and longitude, and is very difficult for amateur photographers to calculate.

Enter the bots.

Fiona showed us astrometry.net. Given a photo of the night sky, the bot tries to identify patterns in the points of light that it recognises from a vast catalogue of space images. Once it has identified a likely candidate pattern, it then examines the rest of the photograph to see if the other features match what it would expect to see. Computers doing clever things with space and time - I couldn’t have been more impressed!

This video is a presentation Fiona gave on the theme at Webstock 2009.

However, even I have to grudgingly admit, machines and algorithms aren’t the answer to everything - yet - and Fiona talked about projects that needed the input of “many human eyes”. Galaxy Zoo led to the discovery of a new type of cosmic object, because participants were seeing a phenomena that did not fit into the rigid classification the project offered them. They ended up doing what Fiona says is now termed “forum based science”. Scientists may be unwilling to join in and speculate, but where there is anomalous data communities quickly form in forums to cross-check and look for other examples. Not bad for the humble web forum, eh? A form of internet communication that has been declared dead many over times...

The social elements of these projects are very important. Fiona said that people will share their findings using any number of hacks like screenshots or cutting-and-pasting data & URLs. They’ve now concentrated on making everything into a “socially shareable object” - and ensuring that any tweets they automatically produce pass “the awesome test” i.e. are they short enough that someone can easily retweet them and add the word “awesome!”

Away from the glamour and beauty of space, another project Fiona talked about was oldweather.org. This is a crowd-sourcing attempt to extract all of the meteorological observations from the museum’s collection of dry and dusty old ship’s logs. The point is to help scientists today map climate behaviour and improve their computer models, by increasing the available historical data.

I did ask Fiona how much resource they had needed to put into community management - it doesn’t take long on the Guardian website for any climate discussion to descend into bickering. Fiona said that actually they hadn’t really experienced many problems at all. I’m assuming this is because whatever orthodoxy the participants adhere to, they are fully expecting the data to vindicate their faith.

Talking overall about “citizen scientists”, Fiona said there had been a big shift in the last few years. “There just aren't enough PhD students to crunch through the available data”, she said, and so releasing large datasets had become the natural thing to do.

She finished by pointing us to one of the latest projects to emerge from a production line that has included Solar stormwatch, a project to spot explosions on the Sun and track them across space to Earth. Just launched is Ancient Lives, a crowd-sourcing exercise to translate the papyri belong to the Egypt Exploration Society.

Keep up to date on my new blog