NGI Code Review Wiki (LEGACY)

For the ethnographers @Leonie and @katejsim

Let’s use this wiki to discuss codes/write coding memos.

If you are discussing a particular code, please make a link to it in the backend. Once the code is resolved, we can delete it from the wiki.


List of codes under discussion and questions for the team go in the wiki. Discussion of the codes itself goes in the comments on the wiki. When the code is resolved, delete it from the wiki. The comments remain to remind us of the choices we have made (and the wiki has edit history in case you need a refresher) .

Codes under discussion:

  1. networking ( and connecting people ( how are we using these in meaningfully different ways? Are they mergeable if we merge them into a code connecting with people?

I initially used the code connecting people to mean the act of connecting two people who otherwise would not have met. We could keep that and also have the code connecting with people or something along those lines? I find networking to be a bit jargon-y.

General coding memos:

On artificial intelligence:

  • artificial intelligence (AI): intelligence demonstrated by machines.

  • artificial general intelligence (AGI): the speculative idea that machine intelligence can understand and perform cognitive tasks as well as or like humans. Use this code for when people are talking about AI speculatively (think Hal-9000)

  • machine learning: mathematical models built on training data to make predictions or decisions. Use this code for when people mention ML specifically or refer to ML-driven technologies like natural language processing/chatbots

  • 'deep learning': type of ML that involves multiple layers of networks, hence “deep.” In single quotes to track this for now. Nested under machine learning.

I have gone through the code: “Collaboration” (code now deleted), which was huge. I broke it down into more nuanced codes, which include:

  • connecting people: bringing people together, interacting with people in a strategic way (this code also contains ‘networking’). I have, for now, opted to stick with ‘connecting people’ and have merged ‘connecting with people’ with this code.
  • connecting ideas: the exchange of thoughts or possible courses of action
  • bringing projects together: connecting or creating a platform for projects, organisers, companies to share ideas and collaborate
  • organising events: planning and facilitating people coming together, e.g. conferences, meetings, concerts, community events.
  • coordinating efforts: people of groups coming together to share resources or work towards a shared goal.

Happy to edit and condense any and all of these codes.


I’m still seeing collaboration in the semantic social network. @alberto, @hugi, any idea why the graph might not be fully up to date?!/dashboard/tagViewFull

Ah, because I only imported the new content and didn’t actually update the graphs. Doing it now.

1 Like

ooh baby new NGI viz looking SHARP! @alberto check it out at k=4

You mean the shift to online neighborhood?

But I still see collaboration:


Because collaboration is still tracked in the dataset, although only once. Those are just the codes that appear with it that single time.

Graphryder Dashboard was not behaving on our main server, so I made some last minute updates.

We are now displaying the NGI data on
This is not running on the main server but on a DigitalOcean droplet where I could actually get it to work. Everything should be working there, including accessing content and seeing the sliders.

Personally I think that trying to get GR to play nice on our main server has surpassed the marginal cost of just operating all the GR dashboards and APIs on another server where it works as expected.


I see multiple co-occurrences with several codes. They cannot be generated by one single occurrence of collaboration. :frowning:


Yes. But hopefully the new GR will be a bit more stable.

I have realized through using GR in Babel that this does not seem to be not how the co-occurrence graph works in the “ego-network” setting. In that setting, each occurrence of a code is counted, meaning that if “collaboration” is coded once and “trust“ is coded twice, that makes for two co-occurrences.

This is only true for the ego-network”, not the full network.

If so, that’s a bug. Let’s remember to create an issue report.

@amelia and team, there is an interesting structural anomaly at k≥5. In the giant component (there are several small ones too, not shown in the picture), one single edge personal data <=> privacy connects two otherwise fully autonomous clusters. The left cluster seems to be about economic sustainability, business models and their social consequences. The right cluster is more concerned with values and the public sphere.

Could this be a promising way of reading the data?

1 Like

for ethno check-in tomorrow

Since joining NGI Forward team in July, I have mostly been coding data about co-working so I wanted to take the time to outline some preliminary findings. As the ego network below shows, we are starting to see some themes cohere around co-working. Since the pandemic, there is a growing demand for co-working spaces, including in rural areas with poor Internet connectivity. Co-working spaces, both formal and informal, are springing up to meet the technological and social needs of working remotely as workers relocate all around the world. Some have had to move for care responsibilities while others are relocating to areas in search of better quality of life. Whatever the reason, in the absence of the workplace providing natural space for social interaction, remote workers are seeking a sense of community. Here’s an excerpt that conveys this:

why customers join our coworking space is, because they want a work environment. Work environments is also connected to community, because if you don’t have more people, you don’t have a work environment, you just have an empty office.

For those operating formalised co-working spaces, their managerial practices of providing safe, clean, and socially distant working environment pose a real difficulty. On one hand, they want their services to last for a long time and see a real opportunity for market growth. However, with talks of the second lockdown and the uncertainty of social distancing guides in some countries, it is difficult to assess whether and how much to invest in operating or expanding facilities (making investment). For the moment, it appears that both the private and public sectors have responded to the demand for communal work spaces, but as the pandemic continues, operators are anticipating different business models for different sectors (distinguishing between sectors).

Some operators are turning to technological solutions like apps to monitor their facilities and regulate members’ access control. This raises some questions about privacy but operators seem to find access control apps to be the most practical option for the moment.

From loss to imagination

Putting co-working in conversation with other semantic nodes (at k=4) begins to tell a complex story about how the shift to online is restructuring social relations:

Following the threads around work from COVID-19 captures the double bind of having the pandemic disrupt the fabric of how we organise our working lives. On one hand, there is profound feeling of loss: the connections stemming from shift to online are public space, social engagement, mental health, and sense of loss. For most people, being unable to access public spaces on their own terms has severely restricted their social interaction. The loss of companionship and feeling of community is contributing to erosion in mental health. This was especially prescient with university students as their social and intellectual lives have had to move online.

At the same time, this sense of loss has also generated new opportunities as social distancing measures are propelling participants to think and act more creatively (imagining the future and imagining alternatives). More on this below.

Sharing spaces, building communities

Another strong connection emerging is on how sharing work space is contributing to a sense of community. Sense of community here captures both what co-working space members’ desire and what operators are doing to bring people together.

Sharing work space means having to share resources and, on some level, sharing, if not negotiating, values. Unsurprisingly, sharing work spaces with social distancing measures is incentivizing many to set boundaries. This code emerged from participants discussing how the burden of maintaining clean space often falls on those with higher standards of cleaning. (I’ve been having some issues with Graphryder on my computer where some connections are sometimes not shown, as is the case with setting boundaries here. Will discuss this at meeting tomorrow). In one recent thread I coded, a co-working space operator talked about the challenges of organising community-building events when everyone is experiencing 'Zoom fatigue'. Here’s an excerpt from one thread:

In the weeks immediately following the shutdown orders, we did jump on and try to make sure that all of our members were participating on Slack, or on our social media channel, and we started hosting a number of Zoom virtual meetings to check in with people and see how they were doing. But after the first few weeks it became pretty clear that most of our members actually weren’t asking for more of those coffee breaks or happy hours. I think people got Zoomed-out really quickly, and we didn’t want to add to that. The way that we were able to support our community was actually by backing off, and so I think that was actually really interesting to see.

How do you bring people together when the one way to connect people is tiring them out? I’ll be keeping track of this dynamic as our participants continue to reflect on the evolving social dynamics of working remotely.


This is excellent. Thank you, @katejsim! Looking forward to your contribution too, @Leonie :slight_smile:

1 Like

@nadia, this part of the writeup might be of particular interest to you (and @MariaEuler + @johncoate) re coworking.

Like @katejsim, I also find interesting co-occurrences when filtering the graph to k=4. For those unfamiliar with our tool, Graph Ryder, this means that we are looking the connections between codes that have been mentioned together by community members at least 4 times.

Let’s start with an overview of the conversation as a whole at k=4. This is everything that everyone is talking about across the NGI Forum.

Screenshot 2020-10-05 at 10.37.33

Screenshot 2020-10-05 at 10.37.01 !

We can see some broad themes emerging as we look at the graph as a whole. Central nodes include codes like personal data, connected to codes like business model and advertising. From this personal data node, we get an idea of some of the issues community members are wrestling with.

If we follow the personal data code to advertising and business model to the left and trade-off to the right, we can see some very interesting debates emerging.
Screenshot 2020-10-05 at 10.40.30

It’s clear even at this level that the use of personal data sets up a trade-off for a lot of participants. In decision-making about sharing or protecting personal data, community members are considering larger context around its use and assessing the impact of the trade-off. cost is also a key consideration. Personal data also connects strongly, if we move south, to agency, in a highly illustrative cluster with control and user experience. We see, again, trade-offs set up – between user control and user experience, raising questions about what one gets when one gives up personal data, and how much control is possible. Community members also raise questions about the effort and cost of controlling personal data in the current environment, when responsibility is highly individualised (individualising responsibility is a salient code at a lower level of co-occurrences).


Moving these abstract concepts into concrete examples, we can look at the left of personal data to see these applied to the specific case of advertising and journalism. Selling personal data for advertising creates revenue for journalism, without which there is an issue of information quality. But there are other potential business models to be considered outside of the selling of personal data, like subscription models. We can also see that community members understand that "free" access to services is not always actually free, since the catch is the selling of their personal data. Yet journalism, without a sustainable business model, is facing some serious funding issues. The fundamental problem is about assigning value to things like journalism, which opens to the question of how to fund it without shady advertising practices. These are all strongly connected to the question of ethics, a key concern in journalism writ large which is closely tied to ethical considerations around advertising, monetisation, personal data, and how to construct a sustainable business model.

Following the graph north, we can see business model and advertising also connected to monetisation and big tech. A series of issues around big tech have been outlined by the community – the need for regulation, the way it perpetuates inequality (facebook is a key example of big tech).
Screenshot 2020-10-05 at 10.49.23

Alternatives to big tech emerge around open source, which has an interesting network built around it – questions of 'empowerment' (does open source empower users?), the need for training to use open source tools, and the question of whether using open source tools opens one up to increased risk. As a technological solution, open source also allows community members to imagine alternatives to the status quo. It also connects to 'human-centred design', as community members ask what exactly it would look like to design digital tools with and for humans.

Screenshot 2020-10-05 at 10.50.39

open source also strongly connects with privacy, an extremely central node in the conversation. privacy connects back to personal data, and we see an illustrative network of privacy concerns articulated by the community: around smart cities and human rights, covid-19 and contact tracing, and as aforementioned, trade-off and decision-making. A salient theme is the question of how to weigh up privacy trade-offs, in order to make optimal decisions about one’s own data privacy. What does it cost? There is uncertainty around how extensive surveillance is, and a distrust of the information that one is given about these technologies, which makes making quality decisions about these issues difficult for community members.


Also connected to ‘trade off’ and assessing impact is a series of environmental concerns. Moving to the bottom-right of the graph, questions of resource consumption and energy are tied to environmental sustainability and cost, showing us a snapshot of the rich conversation emerging on platform around technology, the internet, and environmental sustainability.


Returning to the topic of trust in technology, we see an interesting web of concepts emerging around artificial intelligence. AI is connected to inequality and big tech. It’s also connected to issues of transparency and oversight, as well as effectiveness. This cluster of codes tells us that there is an ongoing conversation around what kinds of oversight of AI might a) be actually effective, rather than performative and b) lead to increased transparency of AI and algorithms. We also see these connected to more concrete questions on what it takes to make these technological infrastructures: their production cycle and the raw materials needed.

@katejsim gives a great analysis of co-working in the post above. I’d like to draw attention to the connections between the internet and online life and the concrete effects those have on life offline. We can see that covid-19 has lead to an increase of working remotely, perhaps unsurprisingly. But it has also lead to a shift of living conditions, particularly co-living and co-working, totally reshaping the division of public space and private space, leading community members to build alternatives and organising space differently.

Part of the impact of covid-19 has been a shift to online and a corresponding sense of loss, of public space, of social engagement, with mental health impacts. One way of addressing this has been to seek a sense of community by reworking offline spaces to increase social interaction. For students and universities, this is especially pressing.

Screenshot 2020-10-05 at 11.00.11

A code I want to keep tracking is of defining terminology, as NGI community members try to cut to the heart of some of the bigger, buzzier issues in tech and the internet. Screenshot 2020-10-05 at 11.06.13

The code is meaningfully connected to nuancing the debate, telling us something about how to move forward and tackle the thornier, more challenging issues in this area in meaningful ways. I’m excited to keep seeing how these conversations unfold.

1 Like

Potentially. Although I think… that the fact that big tech and privacy are on different sides of this might complicate my answer, because I think as we continue to code, big tech and facebook will creep over to the privacy and surveillance side of the graph, potentially disturbing this dichotomy.

It’s also interesting to trace privacy as a code in general, because it’s making us connect two things we often consider separate — privacy on the internet and privacy in real life. I think we are going to end up writing something incredibly interesting on the way that people configure notions of privacy, how it relates to private and public space, and what kind of ‘space’ the internet is. Privacy and autonomy also connect to building communities, both offline and online.

I actually think @matthias’ post on work-life balance vs work-life integration is going to be super interesting as a way in to thinking about this — there’s already great research on the disappearance of work-life separation, mediated by the dissolution of public-private boundaries enabled by the internet, and this is way pre-covid.

The great thing about anthropology is that it doesn’t take these changes as negative (just like @matthias, who sees it as enabling him to live his life in a more integrated way that matches his own personal rhythm). The way this interfaces with living and working with others, tacking between autonomy and shared space, sense of community I think is going to be absolutely fascinating. I also love the connections with imagining the future and building alternatives… the next generation internet is about more than just the online world, for sure, and I think what we bring to bear on it is this question of what it does to communities, how we make them, and how we imagine other worlds together, through our everyday practices.


@amelia Wasn’t able to add quotes in your report post for some reason? Added the quotes directly to my memo here instead.