Workshop Participants "Methodologies for Code in Ethnographic Research"

In my PhD research I approach mathematical modeling and computer simulation of human-environment relations from anthropological Science and Technology Studies as situated socio-material practice with a particular set of ontological, epistemological and ethical commitments in which models as well as modelers are agents, producing and constraining each other (cf. Haraway 1988, Barad 2012, Beck 2008). My central research question is: How are human-environment-relations understood, conceptualized and done in modeling? While modeling biophysical dynamics has made remarkable progress over the last thirty years, the modeling of human agency, social systems and processes in the context of social-ecological transformations is only just beginning. So far, I am conducting comparative ethnographic fieldwork with two research groups using modeling as their central method of inquiry, but differing in their approaches to human-environment dynamics and modeling methods. The comparative aspect of my project aims at operationalizing the notion of situated knowledge specifically in relation to models.

(with Dawid Kasprowicz, not in picture)

We focus on coding practices for scientific research projects. As an interdisciplinary team, we gather together scholars from philosophy of science, physics, and computer science to understand how the usage of code in science is changing the way scientists work. The main interest here is the transformation of scientific models into program code, how scientific theorems are encoded and the numerous ways in which the encoded scientific knowledge retroacts on the models. To tackle these questions, we try to analyze the multiple layers of written code (comments, class inheritances, dependencies, e.g.) through the development and usage of software-tools. Thus, our approach connects to questions from code-ethnography, philosophy of science and software engineering.

Challenges I face in researching code ethnographically: Our aim is to develop analysis tools to present scientific code in a format readable for philosophers, sociologists, ethnologists and other fields of research who do not use code in their everyday research. Regarding the vast landscape of computer languages, we need to find ways to identify meaningful structures across different syntaxes. Furthermore, the way people code can be very different: There are radically different ways to encode one and the same concept and there are multiple ways to solve a problem. Scientists are often self-taught programmers with a purely functional objective under various contexts and conventions. Therefore, we cannot rely on the code we analyse to be written in a clean way or to be well documented. Our general challenge is to provide ethnographers and other code-studying scholars with an easy-to-use-interface that helps to access the functionality and genealogy of the code without having coding experience.

Methodological (or theoretical) topics I want to discuss during the workshop: With respect to the workshop we would be happy to include a discussion oriented at one general and one more specific question. The general would be: How does the usage of code change the way researchers use and understand their models? The more specific is: What role do software-tools play in the research practices of code-ethnography? In which way would additional tools for code-analysis restrict or extend the research questions of code ethnographers? We will provide an example, conducting an analysis in real time and demonstrate the adaptability of one of our tools called GICAT (Generalized Isomorphic Code Analysing Tool).

Examples from my research I would be interested in presenting and discussing during the workshop: We are currently working with 2 different coding projects from various fields of research. They are all mainly written in Python which is in the moment one of the most popular coding languages. To better evaluate our analytics we are working closely with the responsible programmers to have feedback on our interpretations. As mentioned, we would present a short demonstration of our tool GICAT and how we conduct our analysis. The two current projects are:

Gempy (https://www.gempy.org/) which is developed at the RWTH-Aachen University by Florian Wellmann and his team in order to fit 3D geological models to a limited set of probes. Their fitting algorithm is based on a probabilistic approach enabling the geologist to evaluate different models on a small set of measurements without relying on large and expensive sets of probes.

xGaltool (https://gitlab.obspm.fr/dmaschmann/xgaltool) developed by Daniel Maschmann (before at RWTH Aachen University now at Paris Observatory), is a python based tool kit with the main objective to provide an easy data access to extra galactic surveys. It provides data handling and basic analysis tools for galaxy observations ranging over multiple wavelengths. Beside fast access to already published works and associated data bases it also provides standard analysis and visualisation tools.

I am Felipe, a PhD student in Science and Technology Policy at the University of Campinas - Brazil and also a computer scientist. In my research, I follow a model of the Amazon Forest that is being developed by a group of ecologists in Brazil to make predictions about how the forest will react to a future climate. Based on ethnography, STS and anthropology literature, I am interested in what notions of nature are being made possible when phenomena are increasingly perceived only through digital data practices, such as modelling and simulation. I am interested in what gets lost in the process of datafication, but also in what socio-material relation reemerges in and with data, when they are moved, transferred, rearranged or repurposed. In my fieldwork, modelling is very much entangled with coding, and more often than not, it appears to be the sole purpose of the laboratory, making the Amazon disappear in the background of daily scientific practices. As there are almost 20 people in the laboratory contributing to this one model, but also doing their individual research, code gets multiplied in different versions very quickly. And over the years, it becomes harder and harder to call this proliferation of versions by a single name, The Model. In the workshop, I would like to discuss how to follow code changes through time and the relations they engender, how code sometimes becomes central in our fieldwork, and how to keep it visible in our ethnographic writing.

My research aims at gaining an insight on how data executives who strive to expand data collection practices in the sector of consumer behavior analysis reflect upon their line of work, the consequences of the developments they are promoting and whether they are facing some kind of ambivalence between what they do for work and how they themselves want their personal data treated. 

Ambivalence is a central concept of my inquiry. Hence, my research entails questions around the ambivalence of data gathering in the professional field and ethical concerns around data surveillance as articulated within private spheres of family. What are the stakes of this ambivalence? And in a broader sense: How does the distance that data executives strive to achieve between professional requirement and ethical concerns affect the broader discourse around ethical data?

I am foregrounding the ethical aspect of data collection practices and therefore anchor my research questions to critical scholarship on Big Data, Data Ethics and Privacy. I also connect the narratives of ethical dilemmas with broader ethical questions around datafication and data analytics.

I am a cultural anthropologist interested in the changing nature of research design with digital extensions, especially in qualitative research. At the moment, I am a visiting professor for computational social science at LMU Munich, but I am a trained anthropologist and worked on potentials of automatization within discourse analysis as well as digital heritage in recent years.

In my research, I am investigating on concrete digital methods and their possible use in ethnography as well as the epistemological changes involved. I therefore am confronted with computer code which is deeply embedded within the practices as well as the narratives in the field of digital methods development in digital humanities and computational social sciences. Code has its own agency here, whether written by the researchers themselves or in scripts and software used – but how can we investigate this agency and the assemblages code is interwoven in methodologically?

Miriam Fahimi, MA BSc is Marie Skłodowska-Curie Fellow with the ITN-ETN Marie Curie Training Network „NoBIAS – Artificial Intelligence without Bias“, funded by the EU through Horizon 2020 at the Digital Age Research Center (D!ARC), University of Klagenfurt. She is also a PhD candidate in Science and Technology Studies at the University of Klagenfurt, supervised by Katharina Kinder-Kurlanda. Her research interests include algorithmic fairness, philosophy of science, science and technology studies, and feminist theory.

(Miriam will replace Katharina Kinder-Kurlanda, who is not able to attend the workshop.)

Comparative Media Studies / Comparative Literature / Aesthetics of Code

Challenges I face in researching code ethnographically: The lack of documentation of intent when presented with source files, and a difficulty in tracing all the different strands and voices of programmers as layer chronologically; given a particular source file. Who worked on what, to which extent, and inspired by whom?

Methodological (or theoretical) topics I want to discuss during the workshop: I would focus mostly on a literary and stylistic framework, as an indicator of epistemological styles of programming as relationships to knowledge.

Examples from my research I would be interested in presenting and discussing during the workshop: I have a gathered a corpus of idiosyncratic source code excerpts, from code poetry to demoscenes and one-liners which I would be happy to share as case studies of using literary and stylistic analyses to identify the stances of the authors.

Examples from my research I would be interested in presenting and discussing during the workshop: I would like to present some esxtracts from our case studies. We conducted expert interviews with software developers of three tools: Sciscore, Barzooka, ODDPud. And videotaped demonstration sessions in which the typical user scenarios of the tools are presented by the developers.

I’m an assistant professor in Anthropology and associated with the Centre for Social Data Science at the University of Copenhagen. My PhD looked at the norms and values produced through computer science education in Singapore. My current research interests focus on the aspirations for the future produced through tech entrepreneurship. In relation to code, specifically, I've written about how heteronormative binary gender norms are embedded in the ways that code and certain algorithms are taught and understood in undergraduate CS education and more broadly, as well as coding practices relating to ethics, aesthetics, and play.

Ethnographic research of predictive analytics/policing.

Challenges I face in researching code ethnographically: the operators of crime prediction software do not engage with code in their day-to-day business; they themselves mostly do not have access to the code they are working with - however, the code behind the crime prediction algorithms is a key actor/actant in predictive policing practices, which we need to engage with in our analysis.

Methodological (or theoretical) topics I want to discuss during the workshop: How to study code when it is not relevant in the daily routine of the analysed subjects? And in general: How to get access to code after all and how to "observe" it?

Examples from my research I would be interested in presenting and discussing during the workshop: I could present my ethnographic research on the first generation of crime prediction software in Germany (especially referring to the software PRECOBS).