Michael Dyer is part of our Pulse Lab Jakarta family. He joined the team as a geographic information system (GIS) officer in 2019, where he engaged in several humanitarian and development projects focused on the use of non-conventional data to gain higher spatio-temporal coverage. Mikey, as he’s better known, has here offered some personal reflections on his work in Indonesia and the Pacific.
I am a geospatial and environmental scientist, who is passionate about exploring how data and analytics can improve decision making and implementation workflows towards a sustainable future. Over the last few years, I have worked with international government organisations, including the United Nations, as an implementation and geospatial information systems officer across the Asia-Pacific region, particularly in Fiji, Samoa, and most recently in Indonesia. Whilst my time with Pulse Lab Jakarta in Indonesia began around the middle of last year, I was already familiar with aspects of the Lab’s work before then.
My first encounter with the Lab was in 2018, when I met several members of the team in the Pacific. At the moment, they were conducting exploratory research related to disaster risk reduction through the use of mobile network data. I grew particularly interested in the work they were doing, for instance with harnessing mobile positioning data to understand human mobility before and after natural disasters, and therefore kept in touch with the team and exchanged updates here and there. As I learned about the various projects Pulse Lab Jakarta was undertaking, I saw several opportunities in which lessons from the Pacific on the use of geospatial and other data types could be applied to Indonesia, and vice versa. As luck would have it, through an alignment of skills and interests, I found myself soon after working in Jakarta. Below are some insights I came away with; they are not so much about in depth use cases but more on the future direction of GIS applications.
Problem-Solving Through A Spatial Lens
Contributing to regional and global research agendas has been one of PLJ’s underlying functions, which is something they’ve done well over the years by sharing learnings from Indonesia with the broader Asia-Pacific region, and vice versa. When the Central Sulawesi earthquake hit Indonesia in 2018, applying lessons from modelling disaster-induced displacement in the Pacific, PLJ’s team developed a platform to better understand population movements following the disaster. Mapping its data journey since its early days until now, the Lab has noticed increased demand and application of geospatial and remotely sensed data.
Working in Pulse Lab Jakarta and across the Asia-Pacific, I have constantly seen the enormous value of geospatial information, particularly in helping decision makers to comprehend the dynamics between spatial information and public policy. The power of geospatial data lies within spatial information itself, because nearly every piece of data gathered, has some degree of spatial information, which can be utilised to understand and address a range of developmental problems. Spatial data can also complement other types of data, to thus enhance analysis and display contextual information in ways that make sense to different audiences.
One of the projects I worked on at the Lab sought to visualise clusters of areas throughout Indonesia that are dedicated to rice production. Seen in the images immediately below, I was able to do so by overlapping a basic map of Indonesia’s administrative boundaries with data from the 2013 agricultural census. Whilst these visualisations are basic, they served the intended purpose of helping to construct a basic narrative about rice production across the Indonesian archipelago. In addition, these maps can be used as part of early concepting for the development of interactive dashboards and prototypes, for example the interactive analysis and visualisation dashboard we designed using census data to identify smallholder farmers. The potential to understand problems, development-related and otherwise, through a spatial lens is the reason I find GIS exciting.
On the Horizon
As thinkers and innovators in the data and development space, PLJ often looks outward to identify latest trends, emerging datasets and analytical approaches and that’s something I also pride myself on doing. In both Indonesia and the broader Asia-Pacific, geospatial data will play an increasing role in the fields of data science and development. With the growing demand for advanced data analytics and use of alternative data sources in the public sector, more governments are seeing the benefits of exploring remotely sensed data and GIS applications, especially as spatial data becomes more available and more affordable.
During my time in both the Pacific and Indonesia, some of the questions I was commonly asked by government and development partners were:
- Can you automatically detect if a building is damaged after a disaster using satellite data?
- Can you automatically count how many buildings or settlements exist in a particular area?
- Can you automatically detect air pollution or deforestation?
- Can you estimate agricultural yields and productivity with imagery?
To all of these questions, the short answer is yes, however, the approaches necessary to do so are more complicated. For example, there is usually a significant amount of data pre-processing needing specialist knowledge, and remotely sensed analysis typically needs to be ground-truthed. In most cases, different imagery may mean different temporal periods, spatial resolution and spectral signatures, and processing such data demands time and resources. Fortunately, as spatial data and analytical techniques become more accessible and refined, there will be new opportunities for geospatial and remotely sensed data to play a greater role in helping societies and governments work towards a more sustainable future. Furthermore, the adoption of machine learning and artificial intelligence is expected to help overcome challenges related to data processing that are exhaustive, time-consuming or overly complicated to complete manually. Going forward, I believe PLJ will make significant contributions in this space that impact the automation of manual analytical processes and integration of geospatial data to provide added value.
Representation and Challenges
Moving beyond remote sensing data, accessibility to applications that generate spatial data is one of the main differences that I have observed between Asia and the Pacific region. This access is higher in Jakarta and other Asian megacities. When it comes to fintech applications for instance, a huge number of citizens are already using them to access goods and services on a daily basis, and similarly social media platforms are extremely popular across generations. Given the size of the population with these cities, this translates to enormous amounts of data that could be leveraged.
On the other hand, in the Pacific region, financial service applications and applications that provide goods and services are still relatively new and adoption is not as widespread. Essentially, the resources and data techniques being used in Jakarta might not be as applicable among Small Island Developing States (SIDS). However, with the technology sector growing rapidly in the Pacific and the cost of internet accessibility decreasing, there are emerging opportunities for data and geospatial science to play a greater role across the region. PLJ’s shared value partnership with mobile telecom Digicel in the Pacific represents new ways of working together and the benefits of using mobile network data for evidence-based insights.
Across Indonesia and the Pacific, a wealth of traditional data is available, which is typically collected by governments and development actors for certain purposes. In the Pacific, several attempts have been made to centralise access to such data, as the data available is typically scattered across domains and storage units. In Indonesia, the Satu Data initiative which PLJ collaborated with the Executive Office of the President on is a meaningful attempt aimed at strengthening data governance within Indonesia’s public sector. As data availability increases and trust and mutual understanding is built between data owners, data scientists and the public, the underlying potential of traditional data sets and their complementarity with emerging data sets will become more apparent.
Data representation is not uniform across Indonesia nor is it uniform across the broader Asia-Pacific region. Oftentimes, it is the highly populated areas that have a larger data footprint. For example, when visualising data from the Global Biodiversity Information Facility, it is clear that the majority of the 2.3 million records are collected from Western Indonesia and less information is gathered from Eastern Indonesia, as seen in the map below. Whilst I am sure there is information and data out there for these underrepresented regions, it often isn’t available for communities in these locations.
Another challenge in the Asia-Pacific region is that marginalised and vulnerable cohorts often do not have a large data footprint. Therefore we constantly need to ask ourselves questions about the representation of the data. Does everyone in a rural community or a Small Island Developing State have access to reliable internet, a smart device or even a mobile phone? What contextual additional information do we need, if the representation is low? What are some of the privacy issues to expect when working with populations in rural or remote areas? As PLJ expands the focus of its work to Eastern Indonesia and the Pacific, these are some of the points they will want to consider when scoping and designing research projects.
The Way Forward
As data and geospatial scientists in the development space, we need to build trust with our stakeholders and local communities. There were many instances during my time in both Indonesia and the Pacific that reminded me about the importance of this. In particular, I recall developing flood maps in Samoa with a sub-metre resolution, and I was confident in my predictions about which houses and villages would become inundated. However, I had little contextual understanding about what flooding actually meant to persons in these areas, in terms of interruptions, social, economical or otherwise. Critical questions needed to be asked to ensure the information was not just available but useful to them: What would I share with them? How can the information be translated to reduce disaster risk? In other words, showing the communities where flooding was likely to occur simply was not enough. One of the things I admire about PLJ’s approach is that it combines data science with human-centered design to make sense of complex situations, such as exploring how insights from geospatial analysis can be used to improve citizens’ quality of life. It’s a much-needed reminder that humans are at the core of data, and that is the way forward across the Asia-Pacific — interdisciplinary approaches that leverage the skills and assets across data science, social science and other disciplines.
Pulse Lab Jakarta is grateful for the generous support from the Government of Australia