Andrew Su › Likes

@Egon, I'm going to wait for the publication of this article before talking about it :-P - Pierre Lindenbaum

Congrats!! - Björn Brembs

I'm glad you got a paper out, but I'm sorry you couldn't get it published in a real journal. - Bill Hooker

@Pierre... ah, embargoed... obviously increases dissemination of scientific results :) - Egon Willighagen

Asking people whether they care about improving schol comm is like asking whether they approve of truth, justice and ice cream for everyone. Getting them to do more about it than reading a THE article, though -- aye, there's the rub. :-) - Bill Hooker

I should add, lest my previous comment sound overly snarky, that I'm an offender myself: I don't know how many of my pubs are OA, but it's nowhere near 100%. I found a repository that will take my UQ pubs but I left the lab at the end of last century (!) and no longer have preprints... I suck at this just as much as anyone else. (I am resolved never to publish TA again, but that's a different matter.) - Bill Hooker

Actually not sure that caring about improving scholarly comms is motherhood and apple pie. You usually get the "well its got problems but basically its ok....mumble mumble....Churchill...mumble mumble...oh where's the free coffee?". That is even got a response (apparently 10,000 page views in a few days) is much more positive than I expected. Of course who read it, and what their response was is another question entirely... - Cameron Neylon

Anyway, building sympathy has some value even if it doesn't yield immediate action, it lays the ground for later action... - Cameron Neylon

Yes, and that is why I am starting to add peer review and access discussions to my classes, framing it as a "movement" that is gaining speed, and inviting students to consider how their choices vis-a-vis scholarly communication will impact their careers, their patients, as well as their own disciplines. They haven't yet been indoctrinated, and some of them get quite passionate about the topic. - Mickey Schafer

@Cameron, thanks for applying a silver lining of common sense to the dark cloud of my knee-jerk negativity. Absolutely true that getting attention and building sympathy are useful steps and worth celebrating! I didn't actually *intend* any parade-raining there, it was just reflex. - Bill Hooker

It's alright Bill, I always stand well clear of your reflexes... - Cameron Neylon

Nice. Found out in time, this time round - but why don't they ship to Maryland? - Rajarshi Guha

@Rajarshi: Certain states do no longer allow this kind of testing. Just ordered mine. - Josef Scheiber

Just tried that code at that link and it is coming up as invalid - it should still be working on Nov 26, right? - Jean-Claude Bradley

What are the axes? - Bill Hooker

The downward trending curves are the %attendance vs days from the start of the term. The upward trending curves are RSS subscription numbers. At the end of each term the % attendance was in the 10-20% range but the students who attended vs. those who just watched the lectures online performed statistically the same. - Jean-Claude Bradley

Interesting... I wonder what those curves would look like if you _didn't_ provide the screencasts... I imagine that attendance drops in most classes. Or said another way, I wonder if you have enough data to segment the performance of non-attenders who did or didn't watch the screencasts. - Andrew Su

Andrew - I wasn't able to easily identify who watched the lectures and who didn't. When I didn't provide screencasts attendance did drop but not as much - don't have any hard numbers though. Now that I don't repeat lectures and do workshops instead I don't get that curve anymore. - Jean-Claude Bradley

JC, thanks for the info. I really like the idea. One follow up question. Do you ever get feedback from your students that they miss the interactivity of a traditional lecture? For example, suppose a student has a fundamental question at the very beginning of a lesson. In a lecture they could ask for clarification immediately, whereas viewing a screencast would mean they'd have to wait... more... - Andrew Su

This is fascinating. As a terrible swot who never missed a single lecture in three years' undergrad, I'm astonished at the dropoff in attendance. What was the pattern of repeats -- and what does the curve look like with workshops instead of repeats? - Bill Hooker

Andrew - the lack of interactivity in the traditional lecture format was a big motivation for me to change. I simply don't have the time to stop lecturing and spend 30 min on a student question if I have to cover a certain amount of material in class. With 4 hours a week of workshops there is time. Some students watch the lectures with headphones during the workshops and pause to ask questions. - Jean-Claude Bradley

(Oh, and of course, that lady in the right-bottom most advice has a suspicious number of baby deaths during her career... you saying...) - Egon Willighagen

it does say it's about babies. all I can say as the father of twin boys that containing them has numerous options...and they actually anjoy them! - Antony Williams

we've thought about the hosing down one a few times... - Elizabeth Brown

Liz..you've only thought about hosing them....those hot summer days in NCare a perfect excuse for hosing... - Antony Williams

This is today's talk for #rluk10 First time in living memory I've got the slides up in advance! - Cameron Neylon

I may or may not be live streaming at: http://www.livestream.com/cameron... I aim to be live tweeting but that depends on how well tweetdeck scheduling works... - Cameron Neylon

During my Phd I went to the physical library once, only because the printers were located in the basement and I had to get my thesis bound. - Frank

I used to go every Thursday morning. With my pencil and paper cause I couldnt always photocopy the articles .... (and I had email at the time!) oh how times have changed! - Kubke

S4-6 http://friendfeed.com/cameron... - S9 http://friendfeed.com/cameron... - S12 http://friendfeed.com/cameron... - S13... more... - Kubke

S26 http://friendfeed.com/cameron... - S28 http://friendfeed.com/cameron... - S31 http://friendfeed.com/cameron... - S36 http://friendfeed.com/cameron... - S39... more... - Kubke

S54 http://friendfeed.com/cameron... - S57 http://friendfeed.com/cameron... - S60 http://friendfeed.com/cameron... - S65 http://friendfeed.com/cameron... - S67... more... - Kubke

S71 http://friendfeed.com/cameron... - S73 http://friendfeed.com/cameron... - S77 http://friendfeed.com/cameron... - S81 http://friendfeed.com/cameron... - S83... more... - Kubke

S87 http://friendfeed.com/cameron... - S89 http://friendfeed.com/cameron... - S94 http://friendfeed.com/cameron... - S98 http://friendfeed.com/cameron... - - Kubke

Great slides, Cameron. I made very similar points in a talk I gave last spring titled "We are all curators now: Envisioning new roles for research libraries in an era of information abundance, social networks, and digital ephemera" (slides here: http://www.duke.edu/~paulm...) and even quoted you in one of my slides. I also referenced the Shirky quote, and... more... - Paolo Mangiafico

I love slides 59-61. I'm definitely going to steal (and adapt)... - Andrew Su

Like @kubke, I am old. I remember (as in, I actually used) Index Medicus... on paper. - Bill Hooker

Or even save as CSV - Rajarshi Guha

CSV aint much use without the metadata associated with the bizarre directory structure you're using to store the Excel files :-) - Cameron Neylon

And I've got to say, I'm not scared of command line tools but if I want to quickly munge, normalize or check some data, then Excel is usually the fastest way to do it. Maybe I type too slow... - Cameron Neylon

I'm in danger of ranting here. How many times can I click the Like button? - AJCann

heh, not gonna happen. Here's the use case - Researcher has columns of data & wants to a) do stats b) make a graph He throws it into excel, selects the ranges, and gets the result. Next time, his data is different, has different dimension, different labels, different steps in analysis. Almost everything is a once-off, for which Excel is easier, and it stays this way until you're much later in your project, at which point all your data is already in Excel, so you just keep using it. - Mr. Gunn

Mr Gunn I think that's the really key point, the fact that you're doing a different thing every day and you don't start off with a nice collection of scripts to do the processing for you. Not saying it wouldn't be better if you did have them or that it wouldn't be worth building them up but once you start down one path its hard to go down another, esp if no-one around you is. - Cameron Neylon

The other thing that command line people miss is the simple comfort of being able to actually see the data. It also places an interesting limitation on when excel is useful. Once you get beyond a set of things that can be comfortably managed on one screen is when people start to think about shifting to other platforms. - Cameron Neylon

+1 Cam and MrG. Once you get beyond a dataset that you can see on one screen, maybe with a bit of scrolling, then Excel loses some of the comfort factor. But in addition, when you are always doing small-scale datasets, each one different from the last, it's just as easy to fire up Excel. Where I think "those researchers" lose out is that an Excel habit probably limits what you think to... more... - Bill Hooker

I'd *love* to learn more stats and better tools to go with my new mathematical understanding. But I'm not going to, because I simply don't have time. So how about this, data nerds: as well as building me tools, give me some guidelines for making my data palatable to you, so that I can turn to you for help and advice as readily as I now turn to Excel? - Bill Hooker

Provide better tools... like R? Sorry, I do not understand the question. - Egon Willighagen

Egon +1 ... I say this as someone who uses Windows to run one and only one application, Excel, which I like a lot. It's great to summarize some pieces of information, summarize, do simple pivots, but if you want to do any serious data analysis, data mining, pattern recognition, etc, forget it. The people we get frustrated by who do multi-worksheet joins. That's asking for trouble, or if you have 10000 rows. Is there a place for excel, yes, but if you want to do serious data analysis on complex datasets, no. - Deepak Singh

I'm curious too - what else are people going to use? There aren't a lot of tools that are easy to use for entering and processing data. I hope they aren't suggesting researchers use Access instead. - Elizabeth Brown

Even at EBI where Windows is like wearing a sign round your neck saying 'I'm not a proper geek' people have to take account of the fact that Office is one pervasive little suite (well, Word and Excel anyway) and develop for it sometimes (though the ISA infrastructure stuff just apes it, in Java). When I had a real job I used to generate reams of data which I processed with scripts... more... - Chris from twhirl

MySQL is easy .... has some very nice interfaces and can be learned. R is harder, but if you really care about statistical analysis on large scale data, you can learn that. But the better approach, IMO, is to have a software and informatics team that make data and information available to query and visualize. Where a lot of the complex munging has already been done, and you are now... more... - Deepak Singh

Phew - Deepak Singh

Bill H: "Once you get beyond a dataset that you can see on one screen...then Excel loses some of the comfort factor." Interesting. My current project has multiple pages, each 1,305 rows, about 140K cells in all. I was an analyst/programmer for 4 decades. Excel works like a champ for me (with the label row always frozen, of course): What *should* I be using? - Walt Crawford

Deepak's message sounds fine for Big Science. Is there a solution for Small Independent Research? I mean, "software and informatics team..." not gonna happen. - Walt Crawford

Is it a bit early to talk about cloud-based mechanical-turk-y core analysis facilities? Either for the pre-processing (as Deepak describes) or the full service (with coauthorship). - Chris from twhirl

Walt, that's a good point. However, if we train people correctly and get the funding bodies to see the light. Otherwise, we are not in good shape - Deepak Singh

Walt, I'm not the one trying to tell you what you *should* be using. I'm gonna step out of this conversation because it pisses me off to be told I'm doing it wrong, ask how to improve, and be told basically "oh, just use the tools we have spent our careers learning... oh, it should only take you a few hours to learn". Feh. This is called Experts' Myopia, and I suspect I am guilty of it... more... - Bill Hooker

Bill: Good point. I'm not a scientist at all, but even us lay researchers sometimes amass fairly large datasets--and, frankly, Excel is an awfully good tool with a very gentle learning curve. I just wondered what I was missing...and think I know. "Get the funding bodies to see the light" is going to mean $0 for any independent operator. - Walt Crawford

Let me step back a little. I am talking about understanding the ability to choose your clustering algorithm, try and find correlations, figure out time series in multivariate systems. There is a class of analysis for which excel is just fine. There are deeper analytical approaches which require understanding of algos, data models, etc. To be able to proceed without those is like... more... - Deepak Singh

I'm continually impressed by the number of researchers with no command-line or programming experience in my field who are rapidly mastering or trying to learn R. There's a sense in ecology and applied phylogenetics that you'll need R for anything you do, and muddy-boot field biologists fill every workshop offered. I think this kind of motivation + peer training has certainly started to make in-roads... - Carl Boettiger

Carl, that's certainly my hope (from someone who does not know much R) - Deepak Singh

Our bioinformatics core facility does semi-regular classes for members of our institute in R, perl, unix, galaxy, genome browsers, etc. I was actually thinking we should do an Excel class because so many people use it -- if they want to keep using it, they might as well learn it really well. I think part of my job as a bioinformatician is to help biologists learn to use computational tools. - Madelaine

I wonder if someone would be interested in automating the edit/editor counts. The numbers in the old spreadsheet were filled in by hand, but one can imagine writing a simple program using the MediaWiki API to get that info. Might even be a cute application note (BMC Research Notes, PLoS One) that creates a historical and continuously-updated report of wiki activity... - Andrew Su

Thanks for pushing this! Ewen - ewencallaway from email

Ewen -- Hah! you overestimate my ability to "push". Lots of smart people lurking here, so I'm just throwing out a random thought to see if it sticks. Usually an effective filter for good ideas... - Andrew Su

Why would we do free work for Nature? - Bill Hooker

Bill, only if you thought it were personally interesting or beneficial I suppose... - Andrew Su

Eh, sorry, ignore me, that was uncalled-for. The benefits of maintaining such a list are obvious and if Nature is promoting it that's even better. My bad. - Bill Hooker

I agree the list is worth maintaining, but it's just a pain to keep current. (The numbers there now are almost two years old.) Anyway, perhaps a fun little project (for those around here who like that sort of thing)... - Andrew Su

Neil, are you sure a Wiki is the right tool for this? ;-) - Cameron Neylon

Hi, Dan Bolser has kindly created a wiki to collect data on biological wikis: http://nettab.referata.com/wiki..., so if you have anything to add, please put it there! Ewen - ewencallaway from email

Ideally the maintainers of these wikis could supply this info. Perhaps the thing to do is to write/find a mediawiki plugin that would expose this data automatically - then the source wikis would simply have to install the plugin and everything would/might/might sometimes 'just work'. - Benjamin Good