Exploring Big Data to Inform Urban Development Planning

Dr. Ibnu Syabri, from Bandung Institute of Technology, describing the nature of spatial data during the recent Research Dive.

Mapping the dynamics of urban living can be challenging for city planners, especially considering citizens increasing mobility in today’s world. With roughly half the population of Indonesia estimated to be residing in urban areas, urban development planning itself covers a broad scope, including ensuring access to health services, building water supply infrastructure and monitoring energy consumption. This demands extensive amount of time and research, and with new data sources on the rise, experimenting with different analytical approaches becomes necessary. Such was the atmosphere at the Lab’s recent Research Dive for Development, which this time around focused on Urban and Regional Development, and in the true spirit of its tradition, it was a packed three days of intense brainstorming and analysis.

A hot topic among development practitioners, urban development planning has long been on the national agenda for countries looking to advance economically, socially and otherwise. And while much progress has been made over the past decades, the achievement of the Sustainable Development Goals (SDGs) is a raison d’être for new research and renewed discourse. Similarly, it has been the muse for this sixth research sprint at the Lab, which convened a diverse cadre of domain experts, academics and data analysts.

As reminded by past participants during the first Research Dive Reunion held at the start of the year, despite the similarity in format, no two dives are the same. Each Research Dive features a new theme, new datasets, new participants, and as discussed below in the preliminary findings from this research, there are always new insights. Commencing on March 25 with a briefing on the datasets and concluding with final presentations on March 28, in this blog, we recap what took place and share a few key takeaways.

Objectives, Analysis and Findings

The participants were divided into four research teams, each tasked with taking on a specific aspect of urban and regional development and assigned a dataset. Specifically, the tasks included:

  • Task 1: Designing regional development policy, by analysing social events, news media data and its network based on GDELT (a global news media monitoring platform).
  • Task 2: Assessing the accessibility to (emergency) health facilities in Sumatra, by analysing different datasets including transportation infrastructure distribution and health facility locations.
  • Task 3: Monitoring water access for water supply infrastructure planning, by analysing several datasets including municipal waterworks customer distribution data.
  • Task 4: Inferring energy consumption towards urban development, by combining data on social media activity density and socio-economics statistics.

Nevertheless, as with all research, identifying the research question and coming up with possible hypotheses are parts of the first phase into the foray. So under the guidance of specialist advisors and some of our own data analysts, the participants huddled in their groups and spent hours hacking away and dissecting respective datasets.

Highlighted by many back and forth brainstorming exchanges, the research groups managed to assess the nature of each dataset assigned, examine relevant correlations, test out hypotheses, identify trends and patterns, and develop statistical models based on their observations that might be useful for informing urban and regional development planning. Then on the closing day, they took the floor to discuss their findings, policy recommendations as well as some of the challenges in front of their peers and invited guests.

Below we sum up some of the key findings:

TASK 1: The data from GDELT shows broadly that there is a correlation between newsroom reporting and actual social events on the ground, but in particular the analysis reveals that majority of the news covered by GDELT is related to physical activities such as demonstrations, disasters and crimes — and that different countries have different interests. For instance, a significant number of reports suggest that Malaysia and Singapore have an interest in the Indonesian regions of Sumatra and Kalimantan, while many GDELT reports produced by Australian media are about Bali. Also, the amount of news hashed by GDELT on Indonesia affects the number of tourists visiting Indonesia with different impact levels from one country to another. From a policy perspective, this may be used to identify subject matters for regional cooperation and economic growth in line with the Indonesian Government’s Medium Term Development Plan.

TASK 2: Using road network maps, population distribution data, hospital/health centre distribution data, shelter locations, a digital elevation model and fire hotspot distribution, a series of analyses assessed the levels of exposure to two disasters of interest (flooding and haze) in Pekanbaru. The team also examined accessibility to health and potential evacuation facilities, for example the distance to main roads and distance to the city/district areas. The results can be used to propose ideal sites for shelter construction, as well as to map road networks that are either missing or need to be improved.

TASK 3: The research team first derived a few observations from the results of two surveys from the Local Clean Water Company (PDAM) and the Indonesia Infrastructure Initiative (INDII), one of which included the fact that people in Northern Pontianak, West Kalimantan have limited access to clean water (and many of them have only one water source, while people with relatively higher income living in the southern regions have multiple water sources). Based on such observations, the team members tried to propose a potential solution with supporting evidences, which includes a list of alternative locations for water bank installations, supported by rainfall and population data. They hoped that combined with information from socio-economic surveys, these insights can help the Government to improve pipeline water access for low income communities.

TASK 4: Analysing aggregated Twitter data, the research team observed a significant correlation between the energy consumption by districts and the daily activities of Twitter users. Additionally, there was a significant statistical relationship between the amount of tourist activities and energy consumption, so with those observations as well as the socio-economic variance within the population of Bali Province, the team built a statistical model that can infer the daily electricity consumption in Bali using daily Twitter activities with an accuracy rate of 95 per cent and higher.

Understanding and planning for urban and regional development is complex and extensive, and as the researchers concluded themselves, digging into the datasets to uncover more problem-solving insights may take months. Accordingly, the underlying aim of Research Dive is not to simply come with done-and-dusted solutions over the three days, but instead the event aims to contribute to the subject matter by enhancing researchers’ familiarity with the datasets; developing proxies and indicators that may be used to track related SDGs; providing data-driven evidence to inform policy makers; and stimulating future research on the topic in question.

The Research Dives are also a useful resource for Pulse Lab Jakarta in identifying domain experts across Indonesian academia who in turn can help Pulse Lab Jakarta with research projects. The forthcoming technical report will explain each of the team findings in more detail, including how the methodologies used may be improved, how other datasets may help to corroborate the findings, and what steps may be needed for policy implementation.

The research teams were particularly grateful for the expert feedback from their advisors, which included Dr. Ibnu Syabri from Bandung Institute of Technology (who helped to connect the dots between spatial analysis and city governments’ new urban agenda); Dr. Hendricus Andy Simarmata from Universitas Indonesia (who demonstrated how integrating Disaster Risk Reduction and Climate Change Adaptation may advance urban planning in Indonesia) and Dr. Lusi Susanti from Andalas University (who explained why examining urban energy consumption is a must for building resilient, Smart Cities in Indonesia). To find out more on the topics of the past Research Dives, check out PLJ Research Dive for Development.

This Research Dive on Urban and Regional Development is part of a series of events leading up to the Indonesia Development Forum organised by Knowledge Sector Initiative that will take place July 10–11. Under the theme “Pathways To Tackle Regional Disparities Across The Archipelago”, there’s a Call for Papers based on a number of sub-themes, including digital economy for regional development, innovations in local governance and development, among others.


Pulse Lab Jakarta is grateful for the generous support from the Government of Australia.