Advancements in satellite imagery have transformed the way governments and organizations learn about changes on the ground. Satellite data’s ability to provide timely map updates has replaced the burden of utilizing field teams to map locations outside regular activities. This includes new settlements from migrating populations, damage caused by war or natural disaster, as well as identifying small villages in remote places.
Despite the significant shift, satellite imagery has one key drawback: the details of on-the-ground realities that facilitate action (e.g., detailed infrastructure assessment, wayfinding, etc.). The Premise platform and its network of global contributors provide an ideal complement to satellite data with a human perspective.
This blog provides two examples of Premise’s work to support the identification of settlements in two separate scenarios, one in Colombia and the other in Nigeria. These examples highlight the diverse nature of how Premise supports satellite imagery data.
Addressing a Humanitarian Crisis One Settlement at a Time in Colombia
Venezuela is currently in the midst of an ongoing humanitarian crisis fueled by persistent social, political, and economic turmoil. As a result, approximately five million refugees have fled the country, which has triggered the single largest external displacement in modern Latin American history and had ripple effects across the region. Since 2014, nearly two million Venezuelans have migrated to Colombia in search of better opportunities.
Once crossing the border, Venezuelan migrants still face enormous challenges. A proliferation of informal settlements have developed across the country over recent years to deal with the influx of people. Yet humanitarian aid organizations have struggled to identify these settlements to deliver the necessary aid. The conditions of rapid settlement construction create situations of “extreme poverty, poor living conditions, unemployment, food insecurity, and health problems” (Thinking Machines). There is a desperate need for a more rapid and efficient way to detect where these locations exist to distribute essential resources to vulnerable populations.
Premise has partnered with iMMAP, an international not-for-profit organization that provides information management services to humanitarian and development organizations, to identify and validate these informal settlements by combining satellite imagery detection and the ground-truth data collection capabilities of the Premise contributor network in Colombia.
By working with the data consultancy Thinking Machines, iMMAP was able to use satellite imagery to compare a single location at two points in time to deduce if a settlement had recently developed there. These locations became location-based tasks for Premise Contributors to validate and record observations.
Premise guided its network of citizen data contributors through a detailed observational monitoring process, using proxy indicators to report on WASH (water and sanitation), food security, living conditions and healthcare access. Premise aggregates the data into dashboards so iMMAP can visualize the information to support responders toward directing aid effectively into vulnerable communities through sectoral coordination.
Over 350 locations were identified as informal settlements across 68 municipalities in Colombia with 90% accuracy. The other 10% were other new developments such as cemeteries, schools or more formal housing developments.
Pairing satellite imagery detection with the citizen-generated data collection has allowed iMMAP to not only better understand the location of these displaced populations but the reality for basic needs and services at the ground level. GPS location records and photographic evidence provided through the Premise platform can further aid in their decision making for effective resource deployment. This is incredibly impactful in the more remote and rural areas of Colombia that are often the hardest to reach.
Premise plans to continue monitoring these locations on a bi-monthly basis to elucidate needs and track changes over time.
Collecting Foundational Data to Support Healthcare Service Delivery in Nigeria
Nigeria’s vast and dynamic landscape includes numerous communities that lie outside the purview of the official map registry. The rural nature of these settlements, i.e., villages or hamlets, makes it difficult to reach communities to deliver health, education and other services.
The use of satellite imagery combined with the technical knowhow of GRID3 has led to a significant improvement in mapping settlements. In the past, incomplete, static lists of settlements collected by various actors were the only resources field teams could utilize to reach these populations.
Today, GRID3 can produce map layers that are much more comprehensive and accurate. However, a key challenge identified by GRID3 is the “on-the-ground experiences of places” that could identify place names and reconcile conflicting data.
The complex planning for health interventions utilized by the Nigerian Ministry of Health relies on location data for these settlements to allocate resources and personnel for delivering services, such as Polio vaccines or presumably, COVID-19 vaccines in the future. For the field staff to reach these communities, the names of locations are critical to assisting field teams in locating the settlements.
Due to the variance in or alternate settlement names, Premise collects multiple submissions. The data science team at Premise is developing a tool to manage the amount of data to reliably output primary and alternate names for each location for ease of use by partners, another key constraint to effective use of the power satellite data.
In coordination with GRID3, Premise Data is leveraging its network of data contributors to help complement the data set by learning the name(s) of unknown settlements. The initial pilot showed that Premise reliably learned the name of settlements as validated by comparing learned names with existing registries and across multiple data submissions. Premise was also able to add alternative names to the settlements.
Premise is now working with GRID3 to expand this work to all of Nigeria with the hope of scaling globally over the next few months.
Conclusion
Premise complements several types of location data sets by filling in gaps or answering additional questions through on-the-ground validation. Beyond the two case studies described above, Premise has additional recent experience supporting satellite imagery, as noted below.
Last year, Premise data supported the development of an AI tool that accurately forecasts various crop yields. Regional satellite imagery was used to target specific fields, and Premise Contributors located, photographed and marked crop observations in these areas. The data was used to train the AI to identify crops and growth patterns when estimating future yields. Read more about it here!
Due to Premise’s diverse network of contributors, we have unique access to individuals living amidst conflict. With this access, Premise is able to support humanitarian aid organizations to learn more about the particular needs of these affected communities. Where borders or danger prevent travel, Premise Contributors in their own communities provide critical data on damaged buildings, bridges, as well as rebuilding progress.
Together with valued partners like GRID3 and iMMAP, among many others, Premise is helping to better understand the landscape where organizations and companies work. This support is valuable both in the content, as well as the efficiency in which data is collected, to make decisions and produce better outcomes.
Interested in learning more about how Premise can help your organization? Email us at [email protected].