In the six weeks since I last wrote a post with a statistics update, the Citizen Scientists have done 60,628 classifications bringing us up to 354,189. That means the project is still classifying around 10,000 images a week which is amazing given BashTheBug was launched back in April. If you recall, the images of the 96-well plates are taken after 7, 10, 14 and 21 days. Days 14 and 21 are complete for all drugs, and days 7 and 10 are complete for some of the antibiotics. The classifications are steadily mounting for the remaining 7 well antibiotics and by my reckoning we are about halfway to completing these too. Don’t worry; once this dataset is complete, I’ll upload another….
Since the last time I gave an update the Citizen Scientists have added nearly 50,000 classifications. I say nearly, because we are at 293,591 classifications so just 6,409 shy of 300,000! Although the front page reports that the project is only 22% complete, we are much, much closer to having some key datasets finished and able to be analysed – see the previous post for more information. Even though we are 15 weeks after the launch of the project, around 10,000 images are still being classified every week (11,430 last week, which is 1,600 per day). So, we should reach 300,000 towards the end of Monday…
To understand this I need to explain where all the images come from. A small part of each sample of Tuberculosis is injected into each well of a 96 well plate; the plate is quite small – a bit bigger than an iPhone. Every well has a specific amount of an antibiotic dried onto the bottom of the plate, except two wells which have no antibiotic at all. There are fourteen different antibiotics on one plate; one with only 4 different dosages, four with 5, seven with 6 and two with 8.
BashTheBug, and its parent project, CRyPTIC, have been featured in an article by GenomeWeb. This puts both projects into the wider scientific context. Note of caution – reading the content on GenomeWeb is not free, however, if you are associated with a University you may automatically have premium access.
Was idly checking how many classifications had been done up to midnight last night to find we’ve just nudged over the quarter-of-a-million (250,033 to be exact). That is amazing, thank you — a few of the datasets have completed and others are well advanced. Together we are reaching a point where we can start to analyse the data and share the results.
Last week’s post showed how a relatively few number of volunteers do a huge number of classifications. By today, the top 0.3% of citizen scientists (the top ten) have done 31,528 classifications between them, which is a staggering 19% of all classifications. To my surprise, BashTheBug is exactly following the Pareto Principle which says that roughly 80% of events come from 20% of the causes. For BashTheBug after five weeks the top 20% of users (727) have done 82% of all classifications.
Nearly a month after the public launch, the Zooniverse citizen scientists have done a staggering 152,816 classifications. That brings us to 24% through this initial dataset and 47% of the way towards completing an important subset of the data that will let us start some initial analysis. Despite, the number of classifications and new volunteers falling week-on-week, the volunteers still did 20,000 classifications last week and 300 people tried BashTheBug for the first time. And the Zooniverse haven’t emailed their users yet.