Thank you. Yes, sir. As already mentioned, my name is Sophie and I work at the Techno
Anthropology Lab in Copenhagen. I'm here to share with you an app that me and colleagues
and local urban planners and local marginalized communities in Copenhagen have developed.
It's called the Urban Belonging app and of course it's open source. But the reason why
I'm really here to talk about it is because I think the process of how we made the app
is really interesting. So, okay. First of all, PhotoVoice is good to know what that even
is and this is kind of a participatory action research methodology that has been developed
especially to find methods and ways of evolving illiterate communities and has been developed
through kind of 60s and 70s and in the 90s it was really kind of stabilized as a method
with a lot of kind of frameworks and publications coming out. That frames this as a method to
involve local communities in kind of contributing their perspectives and their experiences on
particular issues. And it often also involves not just kind of capturing photos and giving
cameras to local residents and communities but also involving those communities in kind
of selecting and highlighting and doing storytelling with those photos. So kind of a longer participatory
process. And to situate kind of where we developed this app, this was a development that came
out of a project called the Urban Belonging project, hence the name of the app. And this
was really a project that set out to kind of map marginalized perspectives on the city.
So it was rooted in kind of urban planning context and asked, you know, how can we involve
different marginalized communities to shed light on different perspectives on the city?
And in the process we need new citizen engagement tools that kind of re-tool citizen science
and maybe challenge some of the problems that we've seen in the field. So the project involves
a range of different research partners and urban planning partners both in Copenhagen
and in Amsterdam. But most importantly for what I'll say today is that it involved these
different community partners that represented the communities we wanted to in the end use the app.
So we had LGBT plus people, people with physical disabilities, homeless folks, ethnic minorities,
people with mental vulnerabilities, deaf people and also international expats. And so what we
wanted to do in the project was to involve these groups in mapping their experiences of belonging
in the city. And we found out that it was really important to kind of involve the groups from
the beginning and even framing and designing the project. And we used, as I'll get back to,
their inputs to frame kind of the specs and the design of the photo voice app that we then created.
And then the communities went out and used that to kind of collect different data points about the
city, which was then kind of interpreted in these community led workshops and it resulted all in
kind of different exhibitions and co-creation of new frameworks of what does it mean to design
cities for belonging. So that's kind of the overall project. But if we hone in on the development
of the app, which is really what I'm here to talk about, as I already kind of signified in the title,
the project really takes inspiration in two things. So on one hand data feminism and on the other hand
design justice. And so design justice, it kind of tells us that even if we're really well-meaning
as designers, you know, we kind of, we often end up having these abstract ideas or ways of
representing community needs. And most importantly, most of the kind of strategies involved creating
these abstractions where the communities are not at the table of the design process. So we think
that we're being kind of inclusive and so on. But at the end of the day, if the communities that we
assume will use a technology are not at the table themselves, they're not fully represented.
And so Costantel Schaak tells us that, you know, involving communities and community members is
kind of the most important thing to do, not just because of justice and kind of fairness, but
also because community members will essentially have knowledge and experiences and perspectives
that we cannot imagine as a kind of homogeneous group of designers. So this is kind of design
justice. And then on the other hand, we have, or supplementary way, we have data feminism, which
tells us that data infrastructures often are, you know, kind of reproducing or
enlargening the biases and the hierarchies that already exist in society, especially when it's a
homogeneous group of privileged people that sit and do the technology design and design of data
infrastructures. So data can be used and often is used to kind of oppress and exclude and extract.
But of course, we can also think of those same infrastructures and practices as opportunities
of kind of liberation and co-creation. So with these kind of jumping off points, we went into a
design process for our app that really involved a lot of different stakeholders. And as you can kind
of see in this visualization, that becomes very messy. So it's not so linear and easy and predictable
as we maybe would often like, but also it yields really interesting results. So as signified in
the top corner here, we had a research team in Copenhagen, STS, digital methods researchers,
we had visual methodologies researchers from Amsterdam. We also had kind of a starting point
which was that we actually built on an open source app that even existed to begin with,
which of course then had built in epistemologies and so on that we needed to kind of work with.
But then we involved kind of planning practitioners, we involved citizen engagement professionals,
so the people in Copenhagen that work with citizen engagement every day. And then importantly also
the community organizations themselves, which are kind of the red dots. And I think even mapping
and being transparent about when do different kind of actors have an input in such a process
can be an important step of kind of situating our tool and being transparent about how it's made.
But to kind of give you a few deep dives into then well what does the tool do and how was it
shaped in its kind of features by these different actors. So first of all kind of the research
team of course started with talking to urban planners and citizen engagement professionals
about why are we not using photo voice more often when we want to involve a local perspective
in urban planning. And certain kind of limitations arise both from these practitioners but also
from the photo voice literature. And one is that you know photo voice often collects a really
small number of photos and have a pretty limited also sample of participants. And that's because
the way it's done right now is really kind of hands on that a researcher will send emails out
to different participants and ask them to go and use the camera in their smartphone to take pictures
and then send them via WhatsApp or email or something to the researcher who then has to sit
there and kind of keep track on the different participants and the different images that are
coming in and they come in in very unstructured formats as well. So that leads to the second limitation
which is that photo voice data tends to be really unstructured. You can imagine that if you just ask
people to send you photos in a text or on email it will be really different what different participants
do and how much kind of captions and annotations of those images they send. So there's kind of an
issue there that makes it very difficult to actually kind of scale and to use as a stable method.
Then data is also not so safely stored when we ask people to just kind of send these images in you
know text messages or whatever. Then because of this kind of handheld approach it also means
that most projects they tend to focus on one marginalized group because it's really difficult
to manage a lot of different users within different kind of communities and so on. And that is of
course a huge limitation if we want to kind of bring different groups together around discussing
issues at an kind of urban scale. And then finally because it's again pretty handheld the way it has
been done so far it also means that researchers tend to tell the participants where they should go
to capture photos because you need to kind of know geographically where the data is coming from
and since you're not really tracking that at a systematic scale in terms of geocoding or geotacking
the result is often that you're kind of already framing where it matters to go and look at a
certain phenomena. That's of course kind of something we could tweak or rethink. So the app that we've
developed has these key features so it enables kind of a structured collection of photos where
users can be invited into different groups and tasks and are given a certain prompt that of
course can be customized and then that means that we know in response to what is it that we're
actually getting photos collected. Then the app also geotracks where the photos are taken and the
walks that people are kind of going on while capturing these photos. And then most interestingly
I think for kind of quality quantitative research purposes is that we can ask participants to annotate
their images when they collect them on the spot. So we have no issue with kind of recall
problems or memory problems and we get this structured kind of metadata on our images.
And then to add an even other layer we also can ask participants to react to each other's photos.
And so that means that we can start to get this really kind of rich data where we know how the
person who's an author of a photo feels about it but also how other people that look at that same
photo feel about it. That of course invites as you can imagine network analysis and all these
different kind of analytical opportunities open up. So just to show you kind of an example of what
that looks like this is the very simple interface where you can either go for a walk or you can just
take a single picture then you start a task and you kind of are prompted with whatever is the
prompt defined for the particular group. And here I'm on university Wi-Fi so the little devil wheel
is spinning out of control but then when you take the photo what happens is that you're moved into
this annotation space. So you're asked to also say something about the photo and this can be customized
of course project by project what are the interesting kind of metadata points to collect.
You can have kind of scale questions different categories you can have open answer
other categories and so on. And then you can keep going and you can do this either as a single task
you can go on these kind of walks where you you maybe go for half an hour in an area and you
collect multiple photos and then we both track the individual photos but also the way you move
through the city or the space to train. Okay so those are kind of the core features but then as I
have already mentioned the important thing here was that the communities that we wanted to use the
app were involved in kind of designing the specs for it. So some of the things that we learned in
that process I'll just give you a few examples and I mean the list is really exhaustive but just
kind of to give you some pointers of what this can offer. We had a representative from the homeless
group say that you know while homeless people are used to using smartphones and taking pictures
they're really not kind of trained in the Instagram like aesthetics and visual hierarchies of how to
take certain pictures that look in certain kind of way and so they were really worried
that if we invited multiple groups to deliver photos to this project these kind of visual
hierarchies might mean that certain perspectives are foregrounded more than others if you don't have
kind of visual aesthetic skills in framing and taking photos in a certain way. And a similar thing
was highlighted by the Danish Deaf Association representative who said that you know the deaf
community is a highly visual culture but they were worried that if you want people to kind of
document the challenges and the negative things in the city it's really important that people
don't feel a pressure to take photos in a certain kind of way. And so these kind of inputs informed
decisions to not have any editing or filtering options for instance in the app so when you take
photos on the spot and you cannot edit or change them so the light is the light and the way it
looks is the way it looks and we also had decisions to kind of standardize the photos to a square
format so that there's not like differences between what kinds of phones people have and what are the
ratios of the photos and so on and how good are people at framing. We also kind of left out the
option to upload images from the library so this was a kind of purposeful choice to make it to make
sure that people had to be kind of in situ in the city capture things as they are when they see them
and avoid these visual hierarchies that some people can take a million photos and go home and
edit it and then kind of upload the good ones. This has since changed because we've developed the app
of course to be able to also do other things so now you can indeed upload from a library but
another kind of thing that came up from all these different communities was a course a lot of issues
with trust so they have been involved many times by the city this is not only the case in Copenhagen
but everywhere in the world these types of communities are so tired of being involved in
kind of false processes that don't give them any control over the data that they contribute
and where they are not really sure kind of what the data is being used for and
if they wanted to go and kind of delete the data or pull it out of a data set that's really kind
of difficult for them. So we also designed the app to have kind of different features that would
support trust with the community so one is that participants are anonymous in the app so you
cannot kind of know who has taken which kinds of photos and that's of course really important
as it means that it is in workshops or other kind of formats that participants can choose to say
that they have taken certain photos and unfold them and they can also choose to not do that.
We have an option to invite people with email to the app but we also have an option to just
pre-generate kind of anonymous user IDs and invite people via that. We give participants
control over when and how the tracking the geotracking in the app starts and ends
and we give participants options to always see the data that they've collected and contributed
and to always delete it directly within the app. So these are just kind of some of the features
that came from this involvement but I think they're really important in in securing kind of trust
with the communities. So I'll just show you a few kind of examples of what are the kind of analytical
affordances then that this type of data collection opens for. So using the example of the urban
belonging project we had more than a hundred walks by 33 different participants and we had
1400 and some photos collected and of course as I've said you can geolocate them and see where
they are in the city and how people move. You have the images themselves which can be analyzed in a
number of ways. You can use AI but of course you can also use the annotations that people have
created themselves. We even have the reactions that I mentioned so this map for instance shows
distributions of photos with the sentiment in the middle of the dot coming from the author
themselves but then also other people's sentiments or how they feel when they saw the photo and
reacted to it within the app. This is kind of shown as a circle around and this of course is
really interesting in workshops as you can bring up kind of more contested photos and say well why
why are we disagreeing or agreeing on something. We can open kind of geospatial types of analysis
and see well where are different people taking different kinds of photos and we can even zoom
in on kind of the very local level and look at this kind of route data and see well how are
different people moving in and out of a space. So yeah we're seeing this being used for really
interesting kind of analytical cases. These annotations that people are giving to the photos
also helps to zoom in on certain issues. For instance we could see in our project that photos of urban
nature which is what people have said their photos are about predominantly were associated with a
positive sentiment and so that can kind of inform you know zooming in and filtering your data and
looking specifically at photos of nature if that is what is interesting in your project.
So we've done kind of this development of the app for this project but then since then we have
kind of open sourced it and kept developing a whole toolkit around it to make it as accessible
for as many people as possible. So we now have kind of a web-based interface where you can go
and manage groups and set up the tasks and do different translations of the tasks again to
make it accessible. There's an easy overview and export of the data for people that are not programmers
and this is of course really important to make sure that local NGOs, architects, designers,
all kinds of groups without kind of data science expertise can also use it. And then we have this
kind of interactive data dashboard that is just easy to generate and so this is an example of what
that looks like and every kind of project that uses the app can just kind of easily generate
this kind of interactive way of publishing your results or your data set.
So with all of this said we've been really kind of happy to see that the tool which is hosted at
University of Olbo but it's open source it has been kind of used in different ways so we have a lot
of people who are using kind of our version at the university but we also have different companies
who have grabbed the code and set up their own version of this app which is of course what we
wanted and so we've seen really kind of interesting use cases. Gale Architects who were also a partner
in this project they've set up their own version of the app and are running all kinds of different
projects with that. We have had an early use case of the DVSA which is an NGO in Seattle
use it to map environmental justice issues, artworkers with disabilities is another kind of NGO
with different researchers using it so cultural heritage project in Italy and currently the
University of Copenhagen is using it to map perceptions of nature and we also have kind of
municipalities and policymakers use it in relation to urban development. So this is really positive
but I'm also faced with a few questions that I can share with you as I round off. So one is
well what kind of community-based financial model can we set up to secure the kind of continuous
maintenance and development especially as we want to host a version of this app at the university
that can be used by NGOs and so on that don't have resources to set up their own app.
How do we kind of continue to involve marginalized communities and perspectives in how the app
is tweaked and developed from here and then finally kind of how do we offer up this tool
while warning against techno solutionism so because the tool is developed as an inclusive tool we've
also seen some bad use cases where it has just kind of been used without much thought about
methodology or anything else and it actually ends up really disappointing the communities that it's
supposed to serve. So those are some of my kind of questions that I'm struggling with right now
and that we are working to figure out. Yes.
Great we have a lot of time for questions around maybe seven minutes.
Yeah hi thank you. What came to mind was this whole kind of gift exchange for.
Sorry what? Gift exchange. Oh yeah. Is it then data for use of the system or to go back to answer
your questions can it be that we'll receive vouchers or some kind of I don't know events or food
source or something something else that they the participants get in return for. For participating.
Yes exactly. Yes like a larger framework that could be incorporated into the system. Yeah.
So that was my question like where does the data go. Repeat the question. Well I hope that I understand
it correctly but I think you're asking how can you design kind of measures into the to such a
project that actually also gives something back to the communities that are donating their time
and their data. Is that correct. Yes. Yeah I think that's really important and it's very different
what you can do. I mean there's a lot of rules for instance on the university that we can't just
like pay people but I was participating in this project in Seattle where they could do that.
They could actually pay participants an hourly wage for that time and that's of course like the most
useful if you can do that but if you can't then you can find other things to do. So in our project
for instance we did professional headshot photos of all the participants both to kind of use if
they gave permission for that in our report on the project but also for them to use professionally
and that was actually pretty popular and so we tried to find we gave movie tickets and all these
other things right. And I think depending on where you are in the world food kind of vouchers
and stuff like that can be really helpful. Yeah.
Kind of a not fun question but one thing that was lacking in the presentation which was amazing
great is what about the content moderation on the photos? Content moderation? Yeah well that
is the sign into the kind of this kind of this kind of interface that we have online.
So you each participant can only see their own photos so I mean that's pretty protected
and then there's the the moment where you if you want to as a project manager to open the photo
the collective collected kind of photo library up to people so they can see and react to each other
before that happens there's a moderation feature so you can in this interface you can go as a
project manager and say you know are there some of these photos that shouldn't be circulated out to
everyone and of course as a project manager you can also choose to ask participants to go and
review their own photos and delete the ones that they wouldn't have others look at. So it's giving
kind of the tools to the project manager and the participants to do this. Yeah.
I think this is really important kind of point from data feminism is this idea of
paradox of exposure so this kind of sensitivity and awareness that creating visibility is not
just always positive but can also lead to harm and to kind of negative outcomes and so we've really
thought about that a lot and I think it's really important to design for that
both in the tool but also in the process of how you actually use it. Yeah.
Oh there was one there.
Positive neutral or negative. Did you give any thought about like the difference between a photo
being truly neutral versus there's good and there's bad and I see both and so I would classify as neutral.
Is there any way to like get this information out of a picture or is there something that
someone would need to go and like examine after we have something?
Yeah. Well that's a complicated thing to answer in a very short way but the question is how can you
distinguish between positive negative and kind of neutral versus well ambiguous that there's
both positive and negative in the same photo and in our project the urban long project we had the
middle category be kind of ambiguous so that was how we defined it and other people maybe choose to
say that the middle category is neutral or whatever I think that really depends on context
but I also think that what we did here was create an app that can that can generate this really
enriched data set but our entire process was set up around having participatory workshops
so we brought this data to people we brought the photos we brought the maps and so on and then
asked them to tell us and so really the answer is that we cannot answer the question but that we
should bring this data back to people and use the visual nature of it to engage dialogue
and so it might be both ways really but the participants can say so.
Yeah.
Sorry what? The policy makers did they show any interest?
Yes. Yes the cities and city kind of municipalities and city councils and so on they
know that they have a problem like they are lacking tools to engage diverse perspectives
and as we did here multiple groups and not just kind of tokenizing and involving one minority group
so they need this sort of tools and they needed to be structured because that is what gives validity
to a political decision right and that was what was missing before with this kind of handheld
approach to photovoice. Thank you so much.