Software Carpentry

Should We Combine Debriefing and Lesson Discussion?

Tue, 08 Mar 2016 00:00:00 -0600

Belinda Weaver recently posted a proposal for helping instructor trainees finish. While reading her post and writing out the current instructor training workflow for the benefit of GSoC students, I started wondering if it might also make sense to combine the workshop debriefing sessions with the trainee mentoring sessions: a couple of regularly-scheduled slots each week (times chosen to accommodate all comers) would be easier to wrangle, and trainees would get to meet more existing instructors and hear about their in-class experiences. We’d still need some sort of sign-up mechanism to keep numbers manageable, and we’d need someone to report back on which trainees had asked well-informed questions, and it would mean less coupling between the discussion and the specific lesson(s) trainees were working on, but the first two aren’t new work (we’re doing them already) and the third might actually be a good thing, particularly if we follow up on another suggestion and have people attend a discussion before submitting their lesson change. Please add comments to this blog post to let us know what you think of the idea, and to Belinda’s to give her feedback as well.

A Proposal for Helping Instructor Trainees Finish

Tue, 08 Mar 2016 00:00:00 -0600

Instructor training is one way we grow our community worldwide. Yet many people who go through instructor training never go on to teach at a Software Carpentry workshop. Can that be fixed? I suggest that shepherding people though the final stages might help with completion rates.

When the idea of running instructor training in Brisbane was originally mooted, I set up a survey, very early on, to record expressions of interest from people who might want to train. I tweeted that several times in the months leading up to the training. The survey captured name, email address, institution, discipline and how each applicant had heard about the survey. I wish now I had also asked ‘why do you want to train’ or ‘why do you deserve a place over someone else?’ as that would have helped me sift through the candidates.

Nearer the workshop time, I put that survey info into a spreadsheet and got my local Software Carpentry crew to help me whittle the more than 55 responses down to 20. We were mindful of a few things:

having a spread of disciplines, if possible
training people from a range of universities, if possible
training people from outside Brisbane, e.g. from regional universities, or from cities not offering instructor training in this round
training groups rather than individuals (to help foster activity later)
how likely the person was to actually teach.

This last was a bit tricky but attendance at a workshop or having helped at a workshop was evidence of at least some commitment to Software Carpentry.

I think one big problem is that while people like the idea of Software Carpentry workshops, and like being part of that vibe, that community, that doesn’t necessarily translate into a willingness on their part to actually teach. So evidence of previous participation helps identify instructor trainees who have already put in time to make workshops possible – whether by helping organise them, helping out on the day, getting funding to run them, and so on.

But it doesn’t hurt to do some hard questioning beforehand as to whether people can realistically see themselves teaching at workshops, and if so, how often? If someone is a PhD student in the writing-up phase, for example, it’s unlikely they will have much time to commit. By doing that kind of questioning, I was able to eliminate quite a few people from the people who originally volunteered to come.

The 20 people I ended up with were what I considered the best 20. To get a place, they had to complete the prerequisite activities by a deadline, or lose the spot - which is one good way to tell whether or not people are serious. If applicants aren’t prepared to complete tasks, or if they’re sloppy about deadlines, chances are they won’t teach workshops either.

My 20 all did the prep, they turned up on the day, they stuck it out with no falling by the wayside.

After the workshop, we were lucky we had the big Research Bazaar (ResBaz) event coming up at which we offered parallel Software Carpentry workshops in both R and Python. Several fledgling instructors got a chance to teach there while being assisted (and observed) by more experienced instructors. After ResBaz, I organised a Hangout practice session for seven of those recent instructor trainees. During that session, they all taught to the group and they all critiqued one another face to face and via an etherpad. I have since done a second practice teaching session with two other trainees (and walked them both through the pull request requirement) and am planning a third with more trainees from Townsville. I have also followed up twice via email with the rest who attended from Sydney and Canberra.

I think my follow ups have helped people complete, and my ability to get a lot of trainees teaching soon after the training ended focused them on the importance of getting through the final stages. I think having a local ‘shepherd’ can really help get people over the line. But the shepherd could be anywhere - it’s more just having someone specific to whom trainees feel accountable - someone who follows up with them, chases them up, and actually cares that they get the final bits done and qualify as instructors.

Maybe if there is no-one local who could play that role, this could be a job for the mentoring committee to handle - it really is a mentoring task.

The ideal would be to have the instructor trainer hand the class over to the mentor as part of the final session of the training, and for the mentor to check in with trainees regularly from then on – maybe by running teaching practice sessions, handholding, or talking them through the final tasks, getting them over their nerves about actually getting up and teaching (the practice sessions help with that).

I think there also needs to be some contribution pathway for people who’ve completed instructor training but have discovered - belatedly! - that teaching is not for them. In order not to waste their training, they could become an organiser/cheerleader. We could always use more of those!

3rd Annual Big Data in Biology Summer School

Tue, 08 Mar 2016 00:00:00 -0600

3rd Annual Big Data in Biology Summer School

The Center for Computational Biology and Bioinformatics at The University of Texas at Austin is hosting the 3rd Annual Summer School for Big Data in Biology May 23–26, 2016.

The 2016 Summer School offers eleven intensive courses that span general programming, high throughput DNA and RNA sequencing analysis, proteomics, and computational modeling. These courses provides a unique hands-on opportunity to acquire valuable skills directly from experts in the field. Each course will meet for three hours a day for four days (either in the morning or in the afternoon) for a total of twelve hours.

**Click here for more details or to register! **

Great introductory courses:

Introduction to Core Next Generation Sequencing (NGS) Tools
Introduction to Proteomics
Introduction to Python
Introduction to RNA-seq

Bioinformatic courses:

Bash Beyond Basics
Genome Variant Analysis
Machine Learning Methods for Gene Expression Profiling Analysis
Medical Genomics
Metagenomic Analysis of Microbial Communities

Computational Modeling:

Computational Modeling to Study Evolution in Action
Protein Modeling Using Rosetta

New in 2016:

Bash Beyond Basics: This course will focus on being more productive in the Bash shell. We will learn about regular expressions, Unix utilities like cut/sort/join, awk, advanced piping, process substitution, string manipulation, and Bash scripting. Learn to love the command line and increase your productivity with rapid manipulation of bioinformatic data!
Metagenomic Analysis of Microbial Communities: This course surveys the Python software ecosystem and familiarizes participants with cutting-edge data science tools. Topics include interactive computing basics; data preprocessing and cleaning; exploratory data analysis and visualization; and machine learning and predictive modeling.
Clinical Genomics: This course will introduce a selection of genomics methodologies in a clinical and medical context. We will cover genomics data processing and interpretation, quantitative genetics, association between variants and clinical outcomes, cancer genomics, and the ethics/regulatory considerations of developing medical genomics tools for clinicians. The course will have an optional lab component where participants will have the opportunity to explore datasets and learn basic genomics and clinical data analysis.
Computational Modeling to Study Evolution in Action: This course is about the study of evolution using computational model systems. We will use two different systems for digital evolution: Avida and “Markov Gate Networks” exploring many different possibilities of using computational systems for evolution research. Participants will gain a hands-on introduction to the Avida Digital Evolution Research Platform, a popular artificial life system for biological research and the Markov Gate Network modeling framework to study questions pertaining to neuro-evolution, behavior, and artificial intelligence.

**Click here for more details or to register! **

Complexity vs. Subtlety

Sat, 05 Mar 2016 00:00:00 -0600

I gave a lightning talk on Software Carpentry for the OICR yesterday, and in discussion afterward, Jonathan Dursi made an observation that I’ve been thinking about since. He wondered whether the key difference between commercial software and scientific software is complexity versus subtlety. For example, the software that manages workplace insurance payouts for the province of Ontario is complex because it has to handle every regulatory change since the mid-1920s. None of the its rules and exceptions are intellectually taxing, but by the time you turn them into a service, provide a dozen different interfaces for different business roles, and make the whole thing fault tolerant, the software is incredibly tangled.

A lot of scientific software is relatively straightforward by comparison, so long all you look at is the control flow. It’s the specific calculations that are hard: what differencing scheme or statistical test to use, what convergence criteria or significance measure to apply, and so on. And yes, there are a lot of fiendishly tricky algorithms in science, but they’re often hidden in libraries built and maintained by specialists who work and think like software engineers.

All of this brings me back to the issue of testing. (I’ll pause a moment to let long-time readers groan, “Oh no, not this again.”) A lot of tools and techniques for testing mainstream software are really about managing its complexity: some of the most useful books I know about making software right are good precisely because they makes this explicit. Offhand, I can’t think of any good books about managing subtlety—about picking the right calculation to perform rather than handling badly-formatted input data and corner cases in control flow. I suspect this is because subtlety is inherently domain-specific, which means many fewer people know enough to write about any particular bit.

In response to an early draft of this post, Jonathan added, “This distinction is especially important in the early experimental stage of developing a tool: if something is successful enough and widely applicable enough that it becomes ‘hardened’ or ‘productized’ or the like, then the complexity naturally grows to be robust and to handle a wider range of cases.” This is why I enjoy Software Carpentry so much: someone always has new insights. As always, we’d be grateful for yours.

16 February - 3 March, 2016: New Steering Committee, Software Carpentry Value Proposition, Webinar, Vacancies, Community Building, Instructor Training, and Modern Scientific Authoring

Thu, 03 Mar 2016 00:00:00 -0600

Highlights

Our new steering committee has been elected. Congratulations to Rayna Harris, Kate Hertweck, Karin Lagesen, Bill Mills, Raniere Silva, Belinda Weaver, and Jason Williams!
Looking for a good model to show potentially interested stakeholders what value Software Carpentry brings? Jonah Duckles, our executive director, shared his views. Let us know how we can improve on this?

Webinars

DataONE is hosting Greg Wilson for a webinar titled Research Computing Skills for Scientists: Lessons, Challenges, and Opportunities from Software Carpentry on 8 March 2016 at 9am Pacific Time.

New

Over the past three years the rOpenSci community have learnt some interesting lessons about community building and open software. Their paper describing these lessons learnt are now published and may provide some valuable insights to others on similar journeys.
The new instructor training pipeline is starting to ramp up. Five people are now qualified to train instructors while the next round of instructor trainer training is already underway.
We’re currently in the lesson building phase for Modern Scientific Authoring. Comments and contributions are very welcome.

Vacancies

The Bioinformatics Training Facility of the School of Biological Sciences, University of Cambridge is looking for a Training Impact Co-ordinator.

Other

Did you realise that collaborative tools such as the etherpad could be tremendously useful for learners with (even slight) disabilities?
We recently published some statistics on gender representation in the Software Carpentry community in response to a new paper on gender bias in open source. The paper has since received strong critique, but our stats remain.
The first six ELIXIR Software and Data Carpentry instructors recently received their instructor certification. Read more about ELIXR Software and Data Carpentry activities.
18 workshops were run over the past 16 days. For more information about past workshops, please visit our website
Upcoming Workhshops

Communities: The Foundation of Impactful Workshops

Tue, 01 Mar 2016 00:00:00 -0600

I spend a good deal of my time trying to communicate with member organizations about what it is that the Software Carpentry Foundation can do to help them meet their own goals. It is part of my job to showcase to them the return on investment that they’ll see in various areas. The three areas in which this is most apparent are impacts on skills transfer to learners who attend workshops, the lesson material that is publicly available and built by the community, and the capacity building that comes from mentoring instructors who are thinking about impactful instruction in short workshops.

In order to better arrange my ideas I decided to draw a Venn Diagram with circles for Lessons, Learners and Instructors. I’ve tried to use this diagram to build my own mental model for how what we do in terms of core activities can scale and grow. Through this process I’ve come to think of the structure in this diagram as podular (made of semi-autonomous pods), or a fractal element that repeats at various community scales such as university, research network, nation, or worldwide.

At each of our partner organizations we may have lessons, instructors and learners with their own unique and local perspectives, working toward impactful workshops that are appropriate for their own community. Internationally, we are working toward spreading consensus lessons and ethos of using open source tools for open and collaborative science to scientific communities via our workshops while using evidence based teaching methods.

These workshops sit at the nexus of the diagram and showcase, in a focused event, what it is that we stand for. They reflect the impacts we would like to have on changing how science is thought about in the context of computation. The brands “Software Carpentry” and “Data Carpentry” reflect a particular set of opinionated lessons arranged to have specific impacts on learners. This is why we work so hard to make sure that a workshop called “Software Carpentry” or “Data Carpentry” is being taught using our methods (the instructor is “badged”) and with instructors who have studied the community-developed lessons.

Locally when we deliver a workshop we’re working to bring together lessons, learners, and instructors that can deliver impactful workshops. To be prepared for this we strive to convince learners that spending two days with these lessons and our instructors will be impactful and helpful to how they work. When you put this all together, we’re not just developing training and delivering workshops, we as a community collectively own the lessons and are advancing, testing and refining the evidence based teaching best practices that we share with others and reinforce through our instructor training.

Outside of our core lessons and flagship workshops, you are strongly encouraged to duplicate this structure and apply it at smaller scales toward the specific needs of your own communities. This is how we grow and test new ideas and lessons. As you do this we want to know what you’ve learned, we want to hear your success stories, we want to hear about your spectacular failures. Overall, we want the Carpentries together to be a community where the most broadly applicable lessons pertaining to the tools and best practices needed to do modern research can come to be curated and improved together. At the same time, we have thriving global conversations about what gaps there are in our lessons, in our teaching methods, and how can we address those gaps and have more impact on the practice of research supported by computational tools.

One area that this diagram pointed out to me that we could do better at is supporting and helping our learners in their self study. We do know from our website analytics that browsing our lessons is one of the most popular activities among website visitors. We also know, through our instructor survey, that we have many instructors that came to be a part of our community from self study of the lessons online over the years. I would welcome ideas and efforts towards our lessons and our community being more supportive of our learners who are interested in self study. As it is, our lessons are mostly meant to be instructor notes, but if we could find ways to make them more useful for self study, I think that would be fantastic.

What are your thoughts and ideas on this diagram, are there any ways to enhance it or improve it that you see? It has really helped me to organize a jumble of ideas I’ve been dancing around in conversations with partners over the past several months.

Applications due March 1st: 2016 eScience Data Science for Social Good summer program

Fri, 26 Feb 2016 00:00:00 -0600

The University of Washington eScience Institute, in collaboration with Urban@UW and Microsoft, is excited to announce the 2016 Data Science for Social Good (DSSG) summer program. The program brings together data and domain scientists to work on focused, collaborative projects that are designed to impact public policy for social benefit.

Modeled after similar programs at the University of Chicago and Georgia Tech, with elements from our own Data Science Incubator, sixteen DSSG Student Fellows will be selected to work with academic researchers, data scientists, and public stakeholder groups on data-intensive research projects. Graduate students and advanced undergraduates are eligible for these paid positions.

This year’s projects will focus on Urban Science, aiming to understand and extract valuable, actionable information out of data from urban environments across topic areas including public health, sustainable urban planning, crime prevention, education, transportation, and social justice.

For more program details and application information visit: http://escience.washington.edu/get-involved/incubator-programs/data-science-for-social-good/

Bioinformatics Training Impact Coordinator

Mon, 22 Feb 2016 00:00:00 -0600

The Bioinformatics Training Facility of the School of Biological Sciences, University of Cambridge is looking for a Training Impact Co-ordinator. The post-holder will develop (building on existing work), implement, monitor, analyse and report on a comprehensive system of training metrics/key performance indicators across the portfolio of ELIXIR and EXCELERATE bioinformatics training activities. For more information, please see the full ad.

Welcome to the 2016 Steering Committee!

Sat, 20 Feb 2016 00:00:00 -0600

The results are in. Your 2016 steering committee, in alphabetical order by last name is:

Rayna Harris
Kate Hertweck
Karin Lagesen
Bill Mills
Raniere Silva
Belinda Weaver
Jason Williams

Thank you to all of the candidates for standing for election. We look forward to an exciting year of the new committee’s contribution and leadership.

For those interested, the raw results are available here.

More of a Difference Than You Realize

Fri, 19 Feb 2016 00:00:00 -0600

We received this after an online instructor training workshop earlier this week, which reminded me that small differences for some people can be large ones for others:

Thanks for a great workshop the last two days… I wanted to share a separate positive comment that I should have included on the Etherpad: I’m profoundly/severely hard-of-hearing in both ears, and depend quite a bit on lip-reading when listening to people. As such, I have great difficulty with online material if the audio is bad, the speaker is not well lit, or the speaker is simply not on video.

This was the first online course I’ve taken where there were several sites participating, and there was the use of software (Etherpad) for collaborative interaction. I have to admit I was dubious at first at how well this would all work for me with you in one corner, the audience in another, and stuff happening on the etherpad.

In the end, I think it was fantastic. Having everybody collaboratively take notes worked out really well, because then if I didn’t quite get something, I could wait and see if somebody else typed up the information or I could ask about it in the chat window. I did have trouble hearing the audio from some of the other sites, but it wasn’t critical.

Thanks again for a great class.