'Data holds power. But the most important thing is the story you tell with it.'

'Data holds power. But the most important thing is the story you tell with it.'

Researcher, architect and urban designer Saba Golchehr on how data can give a voice to marginalised communities

5 minute read / Rosanna Vitiello & Papawee Sathawarawong

At a time when the smart cities, surveillance and the predictive policing of The Minority Report seem more fact than fiction, researcher Saba Golchehr is something of a social superhero. Championing the democratisation of data mining in the design of our public spaces, she explores how Big Data doesn’t have to become ‘Big Brother’ — but can in fact be the starting point for alternative narratives that give a voice to underrepresented sectors of society. Saba takes a break from her PhD research at London’s Royal College of Art to talk with us about Big Data, shifting powers in local participation, and how social media data can support participatory design.


Saba, what influence has your background had on your approach?
My education was in architecture and urban design and I’ve personally always focused on the human aspect within these disciplines. I'm interested in both the spatial configuration of places and the kind of social interactions they afford. What places are more comfortable to linger in? What spaces do you want to pass through as quickly as possible? Designers can influence those spaces – and with that influence human behaviour.

What is ‘Big Data’? And how does it come into your work?
I recently had this conversation with a computer scientist who now works at the British Museum, Diana Tanase (PhD). She said, "You can interpret data as the artefacts of today’s society." An artefact by itself doesn't really tell you anything if you don't know the context, and the same counts for data.

In my work I’ve chosen to not take a solely arts-based approach towards this topic, instead I’m more interested in the space where design and science meet. There are certain parameters, or data, that one can work with to substantiate a design. In my PhD, I've been looking at how data can inform designers - especially socially engaged designers who empower citizens. I’m interested in how Big Data can help us explore spatial configurations and social spatial systems in new ways.

Do you think data has changed the way we research and understand our spaces?
So far, most data studies that explore the spatial realm haven’t really offered any new or ground-breaking insights. I believe a lot of designers or researchers fall into the trap of letting the data lead the exploration. They become keen on working with new data sources, and therefore only discover what the data allows them to.

For instance, there was a project that looked at mobile phone, bus transport and taxi data to study movement patterns and behaviours of people during a Madonna concert in Rome. Their conclusions basically were, "Well, before the Madonna concert everyone went to the concert hall, and after the concert everyone went home!"

What I'm proposing is that we should try to adapt data-driven approaches in a different manner.

It’s very important to have a question in mind when you’re looking at the data. Start with your question and look at what new ways Big Data enables you to explore it.

That’s interesting - how could that apply to the kinds of socially engaged or community-focused projects you mentioned?
In a lot of socially engaged design interventions, nine times out of ten, once the designer leaves, the whole project disintegrates. I think it has to do with preconceived ideas that the designer has about the needs and wishes of a community. Such designers often also invest a time and effort in creating a new community around their project, while there might already be interested existing communities and social networks in place.

I propose that data could potentially help designers do their homework before they enter the setting.

Instead of having an idea of what the project should be, one can use social media platforms as a digital layer of information. Do an exploration to see if you can find key players in the community. You can approach them and ask what's going on, and they can help you further. It also can tell you a lot about what the community’s interests are, and whether they're engaged in public issues or not; this way you could find out whether the topics you’d like to address are relevant for that specific community.

A few months ago I was approached by Muf architects; they were doing a public realm strategy for Whitechapel. Normally, they would go into the neighbourhood for a long period of time to conduct local observations. This time, they didn't have enough resources to conduct such a study. They asked me if there was a way to do a similar study of local behaviours through data instead. They also mentioned that in their proposal they wanted to advocate for more green public spaces.

I decided to look at Instagram data. I pulled data geotagged ‘Whitechapel’ for the timeframe of one week to see what people’s perceptions of Whitechapel were. Then I did a colour abstraction of those pictures, and ended up with a diagram showing that, at the moment, there is not much green in the area. It's mostly grey buildings, streets, cars. Blue skies, sometimes...

Muf were really happy with this study because these findings could substantiate their argumentation and function in a way as evidence for promoting more green spaces. This is of course not at all a scientific approach, but more of a design approach to data. 

Whitechapel Instagram studies for Muf

Whitechapel Instagram studies for Muf

What about on a city-wide scale? Is data leading us towards smart cities?
The smart city discourse can be categorised within two contexts. The first is cities that are built from scratch - for instance, New Songdo in South Korea - with technology integrated into every aspect of the urban fabric. The second context is all the existing cities that aspire to become ‘smart’. One of the problems for those cities is that it would take a huge investment to place sensors all over the city, and in a few years, those sensors would be outdated and would need replacement. So what a lot of cities are looking at now is how instead they can use sensors that already exist, those that we're all wearing, for instance in our smartphones.

While in the beginning there was a discourse about the ‘smart city’ at an all-encompassing scale, I think a lot of intel companies have also noticed that it's such a huge task to intervene in existing urban fabric. Therefore they now instead focus on local contexts, i.e. neighbourhoods. On this scale there are a lot of developments of ‘civic apps’ where the problem at hand and the local community can be demarcated and therefore can be defined much more clearly.

Where does the non-digital come in? Have you seen successful placemaking projects where there’s an interplay between digital and non-digital?
There’s an example in New York, by a practice called 596 Acres. They started looking at open data from the city council on vacant pieces of land that are publicly owned by the city. They published a website where they showed all this information, but they realised they also need physical presence in the area to activate the local actors. So they printed out several posters with this information and contact details of the relevant agents within the city council. Through this initiative community members got interested, contacted their council, and co-created community owned allotment gardens or playgrounds.

The placemaking practitioners were, I think, involved more as educators of local communities. The practice knew who to address and what people’s rights were, and recognised that these vacant plots could be transformed into valuable community resources, if they could create a critical mass.

© 596 Acres — Poster highlighting the scale of available lots on the ground

© 596 Acres — Poster highlighting the scale of available lots on the ground

What kinds of issues or concerns about data have you encountered?
There are a lot of concerns - amongst the general public, but also within our discipline - about what effects Big Data has on our cities and societies. There is, of course, a big issue with data privacy. All of our data on our actions and transactions is now stored, monitored and used for analytics, which is a scary, gloomy idea. On the other hand, I think that this digital turn at the same time holds a lot of potential to democratise decision-making processes.

The Blockchain is a good example. Rather than a few stakeholders in power, it exists as a distributed network where no one player can dominate. For instance with Bitcoins, every member in the Blockchain has to approve a transaction, before it can go through the system.

I've also been looking at data created by civic technologies - apps like FixMyStreet, or Commonplace, which is a community consultation tool. There's been complaints about people not being able to attend public hearings in these traditional consultation processes. The idea of using technology is that now everyone can access it at any time of day, wherever they want, and they can give their feedback on what they want for their neighbourhood. I studied their data to explore whether this hypothesis is actually valid, which it turned out not to be. Such technologies tend to only empower those that are already in power and exclude marginalised communities even more.

I don't see a utopian view of our future. But I'm also very wary of all of the dystopian approaches.

What I’m proposing in my research is for practitioners to engage with new technologies such as Big Data, but to do so a more nuanced and critical way.

Do you think, then, that those who hold the data hold the power of a place?
Yes. I think data can be a very powerful tool, but the way it is used is very important. Data does speak to decision makers. But it can also shift power relations. I’m very interested in how one can give voice to marginalised communities.

Data can offer an alternative to the predominant narrative.

There is a piece of software, developed by a Silicon Valley start-up called PredPol — short for predictive policing. This software uses historical crime data to predict where the next crime is going to happen. A lot of police departments invested in this software in order to use their resources more efficiently and have their police patrol in the areas where crime would be predicted.

But because police would go to patrol those areas, it also meant that a lot of minor crimes such as selling small quantities of drugs, which would normally be unrecorded, were now on the radar too. This led to a lot more arrests of people for victimless crimes; most of those people come from impoverished neighbourhoods, and most of them from a Black or Hispanic background. The same crimes would be committed by college students in other areas, which are not patrolled, therefore they can get away with it.

So even if the data model tries to be unbiased and colour blind, because of the geographical and cultural segregation in cities - especially in the US - you end up picking out the less fortunate and the marginalised, and putting them into a system of incarceration and re-incarceration. It basically creates a model for punishing poverty.

The Spatial Information Lab of Columbia University and Justice Mapping Center in New York did a very good study on this called Million Dollar Blocks. They looked at the same crime data sets, but instead of focusing on the location of the crime, they looked at where the incarcerated lived, at prison admission rates and the public cost of incarcerating those people. They came up with the term ‘Million Dollar Blocks’: urban blocks where public expenditure would be a million dollars or more a year to put people into the prison system. Their narrative was that this public money could instead be used for better education and other community resources, in order to break the loop of re-incarceration.

This initiative is a very good example of how the same data set is used from a completely different angle. Data holds power. But the most important thing is what story you tell with it. 

© Million Dollar Blocks  — Highlighting prison expenditure per block in Brooklyn, NYC, 2003, represented as dollars

© Million Dollar Blocks  — Highlighting prison expenditure per block in Brooklyn, NYC, 2003, represented as dollars

© Million Dollar Blocks highlighting prison expenditure across two blocks in Brownsville, Brooklyn, NYC, 2003. The data reveals 31 men across two blocks, and 4.4 million dollars spent

© Million Dollar Blocks highlighting prison expenditure across two blocks in Brownsville, Brooklyn, NYC, 2003. The data reveals 31 men across two blocks, and 4.4 million dollars spent


What do you think the role of the storyteller is in relation to big data?

In today’s reality of alternative facts and fake news, the role of the storyteller has regained, I think, a massive importance.

What we can learn from this current predicament is that data or facts are no longer undeniably accepted as the ultimate truth. One important example is the discussion on global warming, where the narrative has gained more importance than scientific facts. News agencies such as the New York Times and social media platforms such as Facebook are also struggling with this reality. Their reaction has been to invest more of their resources in fact-checking.

I think that, in the backdrop of this phenomenon, there is a philosophical discussion about truth — whether truth is something objective that reflects reality precisely, or whether truth is a social construct that can always only be an interpretation of reality. Today we find ourselves in a situation where the latter is more accepted. It's not necessarily a bad thing that we don't see facts as the ultimate truth, rather it pushes us to relearn the art of storytelling while pursuing openness and transparency in communicating our biases.

It's the same for data, because data will always be collected by someone, in a specific way. The analysis will be done in a specific way. Everything is very subjective in the end. Raw data doesn't exist; all data is produced, collected and stored in a certain way, and presented for specific purposes and from a specific view.

The way forward is to be more open about how those investigations are conducted. What kind of analysis is done and what is the agenda behind it? What were the incentives of the people who conducted the research? Communicating the process transparently is as important — if not more — as coming up with new facts.

If you were to give one piece of advice to placemakers in terms of how they could use data, what would it be?
Think about what the values are that you want to achieve and then see how you can find data or evidence to substantiate those values. There’s so much hype around doing something with data nowadays: Big Data is the new thing. I think one rather needs to be concerned about how the data is used and whether it is the right means to that specific end.


We take a walk with Saba around Dalston's Ridley Road Market...

What do you like about Ridley Road?
It's such a lively place - such a cultural mix.

We’ve talked to the local community here about some of the issues they’re facing in demonstrating their value. For instance, one of the shop owners here, who’s Ghanaian, took me into the back of the shop, where there’s a small room and a CD player, and people hanging out. He thinks the place has a really strong social value, but it’s hard to show. How do you think data might be able to help them?

It depends on what he would like to show - if it's about showing that the shop is occupied a lot during the day, you could think of sensors monitoring movement, or having a kind of heat map showing  when and to what extent the space is occupied. But it doesn’t necessarily have to be digital, it can also be talking to shop keepers, asking what their revenue was, or asking the people on the street how often they visit the market and how much money they spend. Such an approach however would take more time, and would result in more qualitative data. But if there is no existing data, it’s also interesting to think about creating your own data sets; what kind of evidence do you want to capture and record? And to what end?

We discussed earlier that analysing data from civic technologies demonstrated that a lot of communities are not represented in local decision-making processes at the moment. I think something similar might be the case with this market: there are all these processes happening; just because they aren't documented in a quantitative way, doesn't mean they don't hold any value.

If you think about the bigger trends of supermarkets becoming more and more optimised and automated - this market could offer the exact opposite. This is where you would go to find social interaction instead. For elderly people, or people who don't have a lot of social contacts anymore, getting their groceries is a way to interact with other people and socialise.

One of the things the shopkeepers have talked about is how they could modernise to bring in a bigger crowd. Could data help to understand who’s not coming to the market?
There is a project in the south of Rotterdam that focuses around markets and how to diversify them. They don't really work with data yet, but I think it's a very interesting context to show the value of these places, how they can compete with large supermarkets and how we can keep them alive.

A key issue seems to be that the council isn’t investing money here.
That happens a lot with council estates. Eventually they say, "Well, these spaces are disintegrated and impoverished and we need to build new housing and new amenities, so we are going to demolish everything.” What concerns me is that the marginalised communities are often unfamiliar with the systems in place, so they don't know their rights and how to fight such interventions. And there's often no one within the council to help them.

Research I’ve read recently about local communities in the UK described members who participate in local decision-making processes as mostly ‘male, pale and stale’! They linked it back to political representativity on the national scale. If you look at the House of Commons, the marginalised communities in the UK aren't represented to the same extent. Because they don't feel represented in the national political arena, I think it also goes back to their local environments, where they feel like they don't have a say or a stake in what's happening around them.

We take Saba to meet Kollier Din Bangura, a local entrepreneur who runs a gallery close to Ridley Road, and is interested in empowering African and Carribbean businesses in the market


We’ve just been down Ridley Road market, and Saba has been talking about data. Do you think we could use data to highlight the value of some of the businesses in the market?

Saba: It's difficult to visualise what is happening in this market through data because there aren't digital transactions. But I think we should invest in a qualitative way of making our own datasets. We used to talk to people and just ask them, on a daily basis, what their revenue was. What's very difficult is to communicate the quality of the market if you can't show it in numbers.

Kollier: Data in terms of footfall? Or financial figures?

Saba: Financial figures would be good, but it doesn't have to be only that. You can have stories of people, which is also interesting. The value of markets like these, where there are social interactions and not just financial ones, are very difficult to quantify. Maybe it's mapping out a network of people.

Kollier: It’s the bait of capitalism. It's all digital.

But here there are things that you can’t quantify — human interaction. I love the service, particularly in the Caribbean shops.
— Kollier

The African shops are very slow. It’s not a machine; there’s just a little more communication and less efficiency. They take their time!

Saba: Yes, and I think that is exactly the added value it has. It's not about digitising this market or making it into the same efficient machine that the rest of the world is becoming.

Kollier: But efficiency-wise, that's the only way it's going to compete.

Saba: What I'm trying to do with my research is show that we shouldn't all be thinking about making the world more efficient, or optimising everything. We think of it from a market perspective and a business perspective...

Kollier: … but not the relationship perspective...

Saba: If you look at the future ideas of how cities are going to be, the social interaction is missing. For instance, apps that are being developed like Citymapper show that you can go from A to B as quickly as possible without disruptions. But is that really what you want? Or do you want to cross through a more qualitative space? Do you want to walk through a park and maybe run into some people?

Where’s that urban serendipity?
— Saba

Kollier:  Yes, and being present in your journey. I've just focused on being present and actually walking around and seeing people. These areas are building so fast and so brutally. How do we create that connection instead? One of the things I love about this area is the personalities, which are lost if you just put in a supermarket.

Saba: That's what happens, if you want every business to end up being the next Sainsburys? From the economic point of view, an economy needs that kind of diversity. And you need to cater to different communities, not just to the general public.


Kollier: I think the key thing is exploration. My challenge is going to be for a diverse audience to actually go into an African or Caribbean shop, beyond the facade. You've got to take your time in terms of rhythm with that individual.

Saba: In the end, people will go to a specific shop for the experience, if there are 10 of the same shops that offer the same products. They will have one favourite shop where they know the person who’s working there. 

There is also this whole turn in the more creative classes where they look for craftsmanship, locally sourced and authentic. You have this market. Why not target that audience?

Local Legends: Yes - like ‘upcycling’ the market. It's also a model where it’s so personal and you do take your time. It’s a luxury environment, like a luxury shop. But you have to think about the difference in the clientele as well, because you don't want to lose the community here.

Kollier: That's why I am looking to empower the younger generation. Keep the information and education about a particular way of doing things, and then ‘upcycle’ it in a way that they can get more business. But the way Africans' sees business is different. Service is so much more than just shopping. That's why it's difficult for someone who's outside of that culture to understand. It's the human quality that I want to quantify.

Saba: One kind of measurement you could look at is time. Then you need the narrative behind that: more time means more qualitative time with the shop owner.

Kollier: Who pays for it? Because time-wise, it won't stack up unless you charge.

Saba: The problem is we live in a neoliberal world where it's all expressed in financial values. Yet having social interaction is how communities really live. Our standard now is that we don't need that interaction. We can just go on with our daily lives without distractions. It's not the way I want to live! It also has to do with having those kind of people in your community, where you can go and have conversations. They are anchors for a local neighbourhood. Without them, the community would disintegrate.


Data speaks volumes to decision makers, as it is often considered to be unequivocal proof of one’s findings. However, data is a tool and not necessarily an objective end-result — the way it is collected is actually subjective. So how we use it to support what we say is quite critical:  A single data set can support multiple sides of a story, depending on how you look at it. “The one with the data has the power. But the most important thing is what story you tell with it.”

Using social media to get a local context
Social media can provide insights into a locale and a community before the design stage begins. It can be invaluable for identifying potential user groups who can act as local agents to help a community gain ownership over a design project. “Do an exploration to see if you can find key players in the community. You can approach them and ask what's going on, and they can help you further.”

Being bound to your curiosity, not to data
When using Big Data often researchers don’t discover new findings; instead they use the data to support something they already knew. But rather than having a preconceived idea,  “It's very important to have a question in mind when you're looking at data. Start with your question and look at what new ways there are to explore it.”

Want more?
Follow Saba’s work at http://www.sabagolchehr.com/ or find her at @SGolchehr