Tag: GitHub

Posted on

Fork, merge and crowd-sourcing data curation

I like to call this one "Fork"

Over the past few weeks there has been a sudden increase in the amount of financial data on scholarly communications in the public domain. This was triggered in large part by the Wellcome Trust releasing data on the prices paid for Article Processing Charges by the institutions it funds. The release of this pretty messy dataset was followed by a substantial effort to clean that data up. This crowd-sourced data curation process has been described by Michelle Brook. Here I want to reflect on the tools that were available to us and how they made some aspects of this collective data curation easy, but also made some other aspects quite hard.

The data started its life as a csv file on Figshare. This is a very frequent starting point. I pulled that dataset and did some cleanup using OpenRefine, a tool I highly recommend as a starting point for any moderate to large dataset, particularly one that has been put together manually. I could use OpenRefine to quickly identify and correct variant publisher and journal name spellings, clean up some of the entries, and also find issues that looked like mistakes. It’s a great tool for doing that initial cleanup, but its a tool for a single user, so once I’d done that work I pushed my cleaned up csv file to github so that others could work with it.

After pushing to github a number of people did exactly what I’d intended and forked the dataset. That is, they took a copy and added it to their own repository. In the case of code people will fork a repository, add to or improve the code, and then make a pull request that gives the original repository owner that there is new code that they might want to merge into their version of the codebase. The success of github has been built on making this process easy, even fun. For data the merge process can get a bit messy but the potential was there for others to do some work and for us to be able to combine it back together.

But github is really only used by people comfortable with command line tools – my thinking was that people would use computational tools to enhance the data. But Theo Andrews had the idea to bring in many more people to manually look at and add to the data. Here an online spreadsheet such as those provided by GoogleDocs that many people can work with is a powerful tool and it was through that adoption of the GDoc that somewhere over 50 people were able to add to the spreadsheet and annotate it to create a high value dataset that allowed the Wellcome Trust to do a much deeper analysis than had previously been the case. The dataset had been forked again, now to a new platform, and this tool enabled what you might call a “social merge” collecting the individual efforts of many people through an easy to use tool.

The interesting thing was that exactly the facilities that made the GDoc attractive for manual crowdsourcing efforts made it very difficult for those of us working with automated tools to contribute effectively. We could take the data and manipulate it, forking again, but if we then pushed that re-worked data back we ran the risk of overwriting what anyone else had done in the meantime. That live online multi-person interaction that works well for people, was actually a problem for computational processing. The interface that makes working with the data easy for people actually created a barrier to automation and a barrier to merging back what others of us were trying to do. [As an aside, yes we could in principle work through the GDocs API but that’s just not the way most of us work doing this kind of data processing].

Crowdsourcing of data collection and curation tends to follow one of two paths. Collection of data is usually done into some form of structured data store, supported by a form that helps the contributor provide the right kind of structure. Tools like EpiCollect provide a means of rapidly building these kinds of projects. At the other end large scale data curation efforts, such as GalaxyZoo, tend to create purpose built interfaces to guide the users through the curation process, again creating structured data. Where there has been less tool building and less big successes are the space in the middle, where messy or incomplete data has been collected and a community wants to enhance it and clean it up. OpenRefine is a great tool, but isn’t collaborative. GDocs is a great collaborative platform but creates barriers to using automated cleanup tools. Github and code repositories are great for supporting the fork, work, and merge back patterns but don’t support direct human interaction with the data.

These issues are part of a broader pattern of issues with the Open Access, Data, and Educational Resources more generally. With the right formats, licensing and distribution mechanisms we’ve become very very good at supporting the fork part of the cycle. People can easily take that content and re-purpose it for their own local needs. What we’re not so good at is providing the mechanisms, both social and technical, to make it easy to contribute those variations, enhancements and new ideas back to the original resources. This is both a harder technical problem and challenging from a social perspective. Giving stuff away, letting people use it is easy because it requires little additional work. Working with people to accept their contributions back in takes time and effort, both often in short supply.

The challenge may be even greater because the means for making one type of contribution easier may make others harder. That certainly felt like the case here. But if we are to reap the benefits of open approaches then we need to do more than just throw things over the fence. We need to find the ways to gather back and integrate all the value that downstream users can add.

Enhanced by Zemanta
Posted on

Github for science? Shouldn’t we perhaps build TCP/IP first?

Mapa mental do TCP/IP
Image via Wikipedia

It’s one of those throw away lines, “Before we can talk about a github for science we really need to sort out a TCP/IP for science”, that’s geeky, sharp, a bit needly and goes down a treat on Twitter. But there is a serious point behind it. And its not intended to be dismissive of the ideas that are swirling around about scholarly communication at the moment either. So it seems worth exploring in a bit more detail.

The line is stolen almost wholesale from John Wilbanks who used it (I think) in the talk he gave at a Science Commons meetup in Redmond a few years back. At the time I think we were awash in “Facebooks for Science” so that was the target but the sentiment holds. As once was the case with Facebook and now is for Github, or Wikipedia, or StackOverflow, the possibilities opened up by these new services and technologies to support a much more efficient and effective research process look amazing. And they are. But you’ve got to be a little careful about taking the analogy too far.

If you look at what these services provide, particularly those that are focused on coding, they deliver commentary and documentation, nearly always in the form of text about code – which is also basically text. The web is very good at transferring text, and code, and data. The stack that delivers this is built on a set of standards, with each layer building on the layer beneath it. StackOverflow and Github are built on a set of services, that in turn sit on top of the web standards of http, which in turn are built on network standards like TCP/IP that control the actual transfer of bits and bytes.

The fundamental stuff of these coding sites and Wikipedia is text, and text is really well supported by the stack of web technologies. Open Source approaches to software development didn’t just develop because of the web, they developed the web so its not surprising that they fit well together. They grew up together and nurtured each other. But the bottom line is that the stack is optimized to transfer the grains of material, text and code, that make up the core of these services.

When we look at research we can see that when we dig down to the granular level it isn’t just made up of text. Sure most research could be represented as text but we don’t have the standardized forms to do this. We don’t have standard granules of research that we can transfer from place to place. This is because its complicated to transfer the stuff of research. I picked on TCP/IP specifically because it is the transfer protocol that supports moving bits and bytes from one place to another. What we need are protocols that support moving the substance of a piece of my research from one place to another.

Work on Research Objects [see also this paper], intended to be self-contained but useable pieces of research is a step in this direction, as are the developing set of workflow tools, that will ultimately allow us to describe and share the process by which we’ve transformed at least some parts of the research process into others. Laboratory recording systems will help us to capture and workflow-ify records of the physical parts of the research process. But until we can agree how to transfer these in a standardized fashion then I think it is premature to talk about Githubs for research.

Now there is a flip side to this, which is that where there are such services that do support the transfer of pieces of the research process we absolutely should be  experimenting with them. But in most cases the type-case itself will do the job. Github is great for sharing research code and some people are doing terrific things with data there as well. But if it does the job for those kinds of things why do we need one for researchers? The scale that the consumer web brings, and the exposure to a much bigger community, is a powerful counter argument to building things ‘just for researchers’. To justify a service focused on a small community you need to have very strong engagement or very specific needs. By the time that a mainstream service has mindshare and researchers are using it, your chances of pulling them away to a new service just for them are very small.

So yes, we should be inspired by the possibilities that these new services open up, and we should absolutely build and experiment but while we are at it can we also focus on the lower levels of the stack?They aren’t as sexy and they probably won’t make anyone rich, but we’ve got to get serious about the underlying mechanisms that will transfer our research in comprehensible packages from one place to another.

We have to think carefully about capturing the context of research and presenting that to the next user. Github works in large part because the people using it know how to use code, can recognize specific languages, and know how to drive it. It’s actually pretty poor for the user who just wants to do something – we’ve had to build up another set of services at different levels, the Python Package Index, tools for making and distributing executables, that help provide the context required for different types of user. This is going to be much, much harder, for all the different types of use we might want to put research to.

But if we can get this right – if we can standardize transfer protocols and build in the context of the research into those ‘packets’ that lets people use it then what we have seen on the wider web will happen naturally. As we build the stack up these services that seem so hard to build at the moment will become as easy today as throwing up a blog, downloading a rubygem, or firing up a machine instance. If we can achieve that then we’ll have much more than a github for research, we’ll have a whole web for research.

There’s nothing new here that wasn’t written some time ago by John Wilbanks and others but it seemed worth repeating. In particular I recommend these posts [1, 2] from John.