We know that relocations after disasters or conflicts are quite challenging. Imagine replicating the shape, the personal bonds, the livelihoods, the urban patterns of a city. So, how to do this without breaking the community but strengthening it?
Embracing the complexity of the urban setting may help to ensure a sustainable and functional resettlement. Our learning is that, this complexity is unfathomable unless you ask the main actor involved, the community. New ICT4D tools, such as network analysis, can help to navigate along this process.
Post-Hurricane Relocation in the Philippines
After the impact of the super typhoon Haiyan, the Anibong community, as many other coastal and informal urban areas of Tacloban, Philippines, was devastated. The government of the Philippines, and different NGOs and humanitarian actors, developed a massive relocation strategy to avoid this calamity to happen again, targeting to relocate 16,000 families into safer areas of Tacloban City.
CRS and the community of Anibong worked together to relocate 900 of these families. Through a community driven approach, the Anibong community participated in the decision-making processes of the resettlement project. This project applied participatory processes to the beneficiary selection, the urban planning of the site, and the housing design, among others.
The purpose of the project, besides creating a resilient community, was to tailor-fit the intervention to each families’ needs, responsibilities and capacities, ensuring a shared ownership, sustainability and suitability.
Using Network Analysis Software
CRS launched a study aiming to match families with plots within the resettlement site, and to assure that the housing type and size met each families’ individual preferences and needs. With regards to social fabric, the study included a simple question to the families:
- Who do you want to live next to? Who are the people that are important for you to be with?
Once the data was collected individually, the project team inputted the data into network analysis software, linking each family to several others, and adding other collected properties to each specific family. The properties included the family size, the potential tenure option, their preferred and minimum plot and house size, their barangay of origin (neighborhoods with administrative autonomy), their livelihoods options, and other specific needs such as disabilities.
The result of the network analysis showed that the families preferred to live within the same 7 barangays of origin. So, the urban planning and location adjusted to this characteristic.
After an individualized matching process of families with plots based in the previously mentioned multiple factors, the entire community verified the result and gave feedback on their plot assignment and location with regards to other families, requesting changes and improvements.
The result can be seen below:
The application of ICT4D technology to understand and handle complexity in urban setting, was an vital tool to ensure a tailored-fit solution for the Anibong Resettlement Project. While it is easy to say, “of course people want to stay within their barangays or neighborhoods,” it can be detrimental to make these types of decisions without further insight and consultation.
The expected results of maintaining the social fabric do not just strive on the physical aspects of the site, but also enabling people to recover quickly by keeping and enhancing the community support structures they need to make this new transition smoothly. It is a lot to ask people to move and start their life over in a new place, network analysis provided a way to lessen this challenge and manage a complex decision making.
Mikel Larraza is an architect and MSc in urban management. He was the former Program Manager of the CRS’ Anibong Resettlement Project in the Philippines.
This application is very interesting but polynode seems to be an app for network/graph analysis, not a neural network. Did they also use a neural network approach for the analysis?
I shared on my company slack and one of our data scientists pointed out that the example given is using network analysis, not a neural network.