For this transatlantic call, we need to be "answering pertinent social sciences and/or humanities research questions". So, I suggest we need to broadly formulate such a question first. One possible avenue is below - let me know your thoughts on this or other suggestions:

The Affective Dimension of Collaboration: Analysis of Structures in Collaborative Development Tool GitHub

In this proposal, we focus on the growing importance of digital collaboration in a global distributed economy. As the global economy switches from centralised mass production to decentralised micro production enabled by digital collaboration platforms, we argue for the importance of a clear, practical and theoretically grounded understanding of effective distributed collaboration, and effective tools to understand how teamwork in a digital, and possibly cross-cultural, context functions and when it is (not) efficient, whom it includes/excludes etc. We focus on GitHub, one of the most popular social platforms for collaborative work, generating millions of transactions daily, etc. We propose the use of sociological theory based in symbolic interactionism as a fundamental framework to enable this theoretical understanding and practical tool development. We are ideally positioned as a team to address these questions, because of our previous successful collaboration both across the Atlantic and across disciplines (psych, soc, SE, AI).

Github is "becoming one of the hottest social platforms for mainstream collaboration"

ADOC: The affective dimension of collaboration. What drives effective collaborations, and what tools can we develop that will promote effectiveness through affective alignment. We hypothesise an important affective component to collaboration that often is ignored or even repudiated. We argue that is should not be ignored, and further that it should be nurtured and promoted because it plays such an important role.

The growing importance of digital collaboration in a global distributed economy - a concept depicting the switch from centralised mass production to decentralised micro production, enabled by digital collaboration platforms, of which GitHub is one if not the most important. If this is where society and the global economy is headed (I read a lot of this stuff recently and could elaborate), we need tools to understand how teamwork in a digital, and possibly cross-cultural, context functions and when it is (not) efficient, whom it includes/excludes etc. We propose that BayesACT is such a tool because it is a valid framework for describing dyadic and small-group interaction, which has the potential to be scaled in order to deal with big behavioural data. However, there are a few challenges, which we aim to address with this project; e.g., how do we identify identity EPA profiles, actions and so forth (basically the questions we have been discussing so far, perhaps a few additional ones). We are ideally suited as a team to address these questions, because of our previous successful collaboration both across the Atlantic and across disciplines (psych, soc, AI). We could enlist your software engineering candidate as well, given his prior experience mining GitHub. This whole idea resonates well with some hypotheses I have been developing while observing the organisational dynamics here at my school. I think there is a lot to be explained about successful organisations and innovativeness based on the idea of affective alignment, but in a “real”/physical organisation it is difficult to observe and operationalise while on Github basically everything is recorded, so we could try to predict the quality/impact/success of projects based on affective properties of the interactions. I am definitely convinced we can frame this in terms of “pertinent social science questions” which we seek to address. How does that sound?

http://dl.acm.org/citation.cfm?id=2145396 http://dl.acm.org/citation.cfm?id=2568315

Some possibly related papers:

• [The Rise and Fall of a Central Contributor: Dynamics of Social Organization and Performance in the Gentoo Community] (http://arxiv.org/pdf/1302.7191.pdf)
• [Are bullies more productive? Empirical study of affectiveness vs. issue fixing time] (http://mcis.polymtl.ca/publications/2015/msrmarco.pdf)
• [All Contributors are Equal; Some Contributors are More Equal than Others] (https://repository.iiitd.edu.in/jspui/bitstream/123456789/388/1/IIITD-TR-2016-001.pdf)
• [Gender bias in open source: Pull request acceptance of women versus men] (https://peerj.com/preprints/1733v1.pdf)
• [The Challenges of Sentiment Detection in the Social Programmer Ecosystem] (http://collab.di.uniba.it/fabio/wp-content/uploads/sites/5/2014/05/p33-novielli.pdf)
• Gender and tenure diversity in GitHub teams

#Work plan sketch

1. Qualitative interview work - what is identity on collaborative platforms such as Github? Analysis of identity dynamics from a qualitative perspective based on a small sample (N=10) 📝
2. Surveys to characterize identity and behaviours and gather demographic information. e.g. how does a pull request feel to different members, etc
3. Choice of specific focus projects, for each:
4. gather data from the group to seed the model 💽
5. run a group BayesACT model simulation 💻
6. compare predictions of the model (e.g. productivity). We expect e.g. groups with better alignment structure + who have differing skills to move forward faster 📈
7. generate more complex, testable hypotheses at the large (group/network) scale (e.g.diversity increases productivity), and at the micro (individual) scale (e.g. person X needs more attention 😢 as in Bayes.social)