Understanding data activism through civic activism in waste pollution

Data activism: waste data for society

As Urmas Paet, Member of the European Parliament, has noted: “Digitalization is here to stay. It is not going away. The best everybody can do is to step in, shape the tide and work for harnessing the digital dividends while making sure no one is left behind, somewhere on the other side of the digital divide.” One can already observe that ICTs are active in almost all development issues and sectors while technological solutions are also increasingly part of all aspects of development processes (UNCTAD, 2014). The question is, how can civil society, a stakeholder in global sustainable development, harness digitalization to deliver better outcomes in their own work. Civil society organizations have the need and the commitment to keep up with the pace of digitalization to be able to properly shape it and use it.

One global challenge millions of people are actively taking action against is mismanaged waste (trash). It is everywhere. Left uncollected, it often ends up harming the environment. To clean it up efficiently, it is important to focus on the places that are most affected. But how do we know where in the world that is? Let’s Do It Foundation has been mobilizing millions of people in cleaning up waste – that includes also mapping unauthorized trash dumping sites at global scale. This mapping has relied primarily on human volunteers reporting dumping sites through their mobile phones. 

The resource intensiveness of relying on humans for identification, assessment and mapping greatly limits its scalability. The techno-enthusiasts also highlight that humans are undependable as sources of information: people lie in surveys and pretend on social media. People are not consistent in their reporting, they might report trash but not report cleaning it up and might not report the cleanliness of the same location. This raises the question of how technologies from augmented reality games to intelligent drones and image assessment might be able to help automate the process of mapping the world’s trash (Leetaru, 2019). 

Let’s Do It Foundation initiated the recent biggest civic action taken against waste –World Cleanup Day. An action that draws attention to the global trash problem and triggers behavioural change in masses. But the long-term goal of Let’s Do It Foundation is to avoid unnecessary waste generation in the first place. That means that only focusing on behavioral change would prove to be insufficient. By combining artificial intelligence with publicly available photos we have developed an AI algorithm WADE (waste detector) for detecting trash in geolocated images. The algorithm is able to survey images all over the world, locating trash on a global scale. This paves way to data activism in waste pollution that helps not only to wisely plan cleaning but also to help respond to waste pollution root causes and plan best remedies. E.g. we may reveal locations where waste regularly appears even if cleaned up repeatedly, indicating that other measures for keeping cleanliness might be more effective. 

Data activism and big data in tackling waste pollution

Data activism could be considered as self-organized, citizen-controlled, self-managed and noncommercial activism, which funnels content that communicates visions of an alternative world for social change utilizing the data infrastructure. For data activism to be effective, tiers have to be climbed to progress from metadata to knowledge, and horizontal steps have to be taken as well, linking analytical to critical knowledge (Gutiérrez, 2018).

The three types of knowledge in waste data could be defined as such: 

• Empirical-analytical – what kind of waste pollution can be encountered, where is it and what are its characteristics. 
• Hermeneutic – interpreting the reasons behind the pollution. The reasons can vary – anything from historical (e.g. we have always done so) to political (e.g. littering as an act of defiance against the government). 
• Critical – perhaps the most important and difficult tire of the data. Understanding the journey of a material requires a thorough understanding of the context – how does anything become waste and end up as pollution? Most work by activists groups is done on this level, not the least to validate the knowledge derived from any big data collected on the subject. 

Let’s Do It Foundation has throughout the years coordinated mapping campaigns to support campaigning for cleanups and plan volunteer allocations. The WADE development and our World Waste Platform have opened new opportunities for more and up to date information to be collected and processed. However, monitoring only gives us the empirical knowledge. Data activism in waste pollution is yet to embrace big data. Not the least because of lack of resources, and in all honesty, a client. The potential is surely there but still very little implemented, or it is unknown to the activist groups treating waste pollution. For example, knowing that neighborhoods that are greener and cleaner are often safer, and using new surprising correlations between cleanliness and level of crime, enable researches use Google Street View to predict crime patterns (How surprising …, 2015). 

To analyse the state of big data (or its absence) in waste data activism, I will look at the seven ‘V’s that characterize the nature of big data (Gupta and Rani, 2019).

Volume – it means there is no well-defined quantity of data and it is constantly increasing. World Cleanup Day, Earth Day Network, Break Free From Plastic and other campaigners are engaging an increasing number of people in their campaigns, including those that map and report waste. Today, most such activities could be called data science or analytics. E.g. producing a report on most polluting brands. However, over time, the datasets grow and more and more information can be connected to the findings, such as correlation with spread of diseases (e.g. due to sewage clogging) simultaneously running advertisement campaigns (e.g. to analyze the impact of a campaign to consumption behavior) or info about the weather conditions (e.g. to predict how waste may travel from one location to another or what kind of weather conditions may impact disposal behavior).

Velocity – Today, we can count around 10 more widely used applications that allow users to map waste and/or report the cleaning of it. These solutions alone unite a community of ca 400 000 users around the world. When mapping waste for World Cleanup Day 2018, a daily count by Let’s Do It Foundation was 121 waste points through five different applications. Litterati, one of the more widely known examples of using mobile technology to lead people where the litter is, receives thousands of data points a day. As a comparison, WADE is capable of analyzing 8 415 360 images in one day and detect both where waste is and where it is not. Whereas most waste mapping communities handle a smaller amount of data, the nature of the data is the same or comparable. As a global community, there is a huge potential to create a big data set that could be used for a more impactful advocacy work and more accurate prognoses of agents of change in communities around the world. Most importantly, Let’s Do It Foundation released its trash mapping algorithm WADE as an open source tool available for free on GitHub. This means that not only can others help contribute to the project, but anyone can now use and adapt the algorithm to their own trash mapping needs.

Variety – Waste related information is often manifold and contains a wealth of information. More than trash can be detected from images. Also the surrounding areas, the ecosystems where trash is or is not, is important. Hashtags can identify brands and products (e.g. #McDonalds, #Plastic). Geotags map problem areas. Timestamps indicate the times of year we see specific types of litter. 

Variability – the data flows are unpredictable. On a movement scale, the kind of data that is collected about waste varies greatly. Info is collected about the type of material of the waste, the brands, locations and environmental impact. Sometimes waste is only reported, other times also the fact of collecting it. The data points are not geographically consistent. Footage may be anything: a recording of high-rise littering like they detect from tall buildings’ windows with surveillance cameras in Singapore or high-way cameras in Nigeria.

Veracity – completeness, accuracy and quality of data. Waste pollution data comes from many different sources, intended for different uses. On one hand this creates a motivation to promote a standard for data collection across different groups. On the other, it provides a challenge for algorithms to learn to link, match, cleanse and transform data across systems. There may be an opportunity to create value for waste pollution data by connecting multiple data linkages.

Value – benefits of data. Until recently Let’s Do It Foundation used globally collected waste data for a social good – campaigning and planning resources for cleanup actions. Possible use cases for automated waste pollution data could be predicting critical waste areas, analyzing how legislation (e.g plastic bag ban) affects the reality or detecting correlations between disease outbreaks. To make an increasingly accurate analysis of waste pollution possible, a global database initiative World Waste Platform was called to life, but it still remains unclear who would be the interested parties able to fund such developments, and more importantly, willing to act as a result of the learnt outcomes. 

Litterati has identified several benefits to analyzing waste data. For brands it can mean becoming more environmentally mindful. It could lead to product innovation, sustainable packaging, and educating customers. Cities can use the data to help pass a tax on sales of most littered items, e.g. San Francisco has generated a multi-million dollar revenue stream from tobacco retailers after a case backed with data was made against them, showing the vast amount of cigarette butts polluting the city’s streets. 

An unexplored business opportunity lies with telecom companies. They are the closest service partners (sometimes the only ones) to individuals with a mobile device. On the one hand they could attract more users through meaningful engagement, on another provide the society data processing service that civil society could never have. 

Validity – correctness of data for intended use. 50 000 data points analysed by WADE have been validated and proven to be 90% accurate. It can serve as an open source tool for academia, policymakers, education and business to see trash and plan solutions.

The above shows there are already many signs or great potential for big data to be used in data activism for waste pollution. Once the information is there, the first tire – the empirical knowledge – should be linked into more actionable critical knowledge.

Turning information into knowledge

Data forms the base or bedrock of a knowledge pyramid: data precedes information, which precedes knowledge, which precedes understanding and wisdom (Adler, 1986; Weinberger, 2011). Each layer of the pyramid is distinguished by a process of distillation (reducing, abstracting, processing, organising, analysing, interpreting, applying) that adds organisation, meaning and value by revealing relationships and truths about the world (Kitching, 2014). The steps in the so-called knowledge pyramid in case of waste data could be defined as follows:

Data – The very first element, the identification of pollution or the absence of it requires analyzing images that could be collected either manually by individual users, with autopilot detection systems like dashboard apps for car drivers or by scanning through public surveillance cameras.

Information – there are many interpretations to what information is but at the very least it would differ from data by enabling to give data some meaning or interpretation. With the case of waste pollution it would mean determining what is waste on an image. In automated waste detection, the complexity of data comes from the fact that not everything that looks like trash is trash. Trash is a word people use for an object that lacks purpose, and the purpose of an object is often not obvious in the images used for teaching an algorithm to spot trash (Let’s Do It AI). 

Knowledge – it is through processing, management and usage that information is converted into the even more valuable knowledge (Kitching, 2014). This could mean addressing the issue by choosing the best mitigation plan, e.g sending a notification to the right authority, generate a petition for the community.  

Wisdom – Once a solution has been chosen, it needs to be validated. Wisdom can be gained by learning that the mitigation plan has been unsuccessfully implemented (e.g. the authorities have not reacted) or if implemented, the wished result was not achieved (e.g. pollution reoccurs despite the chosen mitigation plan). The faster the feedback and thus acquiring wisdom, the more resource effective can systems become in tackling the root causes of pollution. 

In the case of waste pollution, data literacy can determine or define the path an organization would choose to tackle the problem. Data do not exist independently of the ideas, instruments, practices, contexts and knowledges used to generate, process and analyse them (Bowker 2005; Gitelman and Jackson 2013). There are numerous underlying reasons why waste would end up polluting public spaces like care of place, responsibility factors, penalties and rewards, poor infrastructures, inadequate policies or the lack of enforcing them, etc. Often it is many of the reasons combined, which may cause for an advocacy or activist group loose focus. A good strategy must involve analyzing the local community and reflect smartly the data describing the situation.

Another important part of making knowledge applicable is choosing the right way of presenting it – communication and visualization. When it comes to waste pollution, it is difficult to grasp the magnitude of the situation and the personal or cooperative stake in it. It feels different by location, territory or beliefs or may seem irrelevant compared to everyday life. There is a need for common standardized but intuitive method for every individual, culture representative, role or occupation to understand the situation similarly. This has been the underlying reason for the World Waste Platform – to visualize the volume and emphasize the fact that mismanaged waste is a global problem. 

Summary

Individuals are made governable; they display what Foucault called governmentality. Numbers create and can be compared with norms, which are among the gentlest and yet most pervasive forms of power in modern democracies. (Porter 1995: 45). In the race towards sustainable world order, data activism is a powerful vehicle for advocacy groups against corporate resources and seemingly rigid political systems. Although underfunded, the civic movements against waste pollution are rich in unyielding, growing and highly motivated community of globally connected people. The more advanced we become in presenting more quality data and the form of it to policymakers and voters, the higher are the chances for data activists to prevail. We can’t manage the waste pollution if we cannot measure it.