Taking Advantage of Cash Delivery Data Challenge Statement

Challenge presented by Catholic Relief Services


Generate a more objective response analysis at the beginning of new humanitarian crisis, and limit human intervention in the decision of funding of the intervention, intervention strategy and response options, including delivery mechanism.

The Problem

After a crisis, it is difficult to make decision, get the funding, and have an holistic response timely and at scale because of the low-tech data collection combined with consensus-based decision-making process of the coordination mechanisms. In a lot of sectors robots help making better and faster decisions for doctors, lawyers or investors, why not do the same for humanitarian interventions?

What are some possible solution to this challenge, and what are their limitations?

CRS is accumulating a large amount of data from beneficiaries registrations and assessment over the years. Combined with secondary data (data from other humanitarian organization publicly available, satellite imaging, etc.) all of this data could be analyzed objectively by an algorithm and turned into suggested interventions (based on previous experience in the same / similar country).

What has been done so far?

CRS is in the process of standardizing and storing into the cloud its data points into a global data warehouse. CRS also uses API such as the CAT system to perform on the same platform beneficiary management, data management and transfer of humanitarian assistance. CRS is promoting innovative collaborative model with other organizations of the CCD initiative to improve coordination by revisiting and offering more efficiency to the humanitarian value chain, particularly around cash delivery. Moreover, with START, CRS is piloting a parametric insurance initiative in Senegal. Finally, CRS is exploring an area-based approach for coordination and intervention for humanitarian assistance.

Could you list one or more user stories for the most important cases that should be solved by this solution?

  • As a person regularly affected by natural disaster, I want to receive as quickly as possible humanitarian assistance in a “one stop shop” manner so that I am able to resume quickly my activities and be better prepared and resilient for the next crisis
  • As a humanitarian coordinator, I want to be assisted by a roboadvisor to help me identifying needs, market and other infrastructures and services capacity in close to real time so that we can appeal for right amount of funding and adjust our intervention and supply of the humanitarian assistance to the ever changing context of the response.
  • As a financial service provider, small vendor, humanitarian organization -not in clusters system-, local partner, I want to have access to close to real time context and need analysis, and response scenario to be able to provide what people need the most at my community and individual level directly so that I direct my service where it is needed the most
  • For example:
    • As a project manager, I want to know automatically each day which clinic performs below average so that I can intervene right away.
    • As a country director, I want to see all our weekly reporting data on a single website so I can give donors better updates.

Who are the key players (so far or should be engaged in future)?

  • Cloud industry companies to aggregate existing and new data per area of intervention
  • Companies using algorithm for scientific-modelling based scenario used for forecasting (weather, epidemic, conflict, etc…).
  • Technology solution provider to gather individual data (heart rates, etc…) and community level data (e.g. water level)
  • “old school” humanitarian responders to respond where the technology is not available, or to “correct” decision suggested by roboadvisor because of the good knowledge of area of intervention and qualitative approach
  • Government for regulatory environment and decision making on strategy of intervention (sovereignty of government in natural disaster is important. However might be different in context setting, or at least will requires better data protection/beneficiary privacy to not compromise HLP)
  • Fintech, FSP as solutions will be more monetized
  • Insurance companies, investors (see Insurance Linked Securities model such as parametric insurance and catbonds)
  • Blockchain industry providing digital ID and smart contracting
  • A bunch of lawyers and ethics committee to solve new problems these solutions will bring

What are the possibilities you see for better leveraging emerging data science & AI capabilities?

  • Sandbox testing with existing (huge) data set available, particularly in protracted crisis.
  • Micro testing in small area of intervention where technology is available (urban crisis)
  • Build upon technology and experience existing in finance, insurance, medical and law sectors on roboadvisor.
  • Standardizing data points (e.g. HXL langage) for technical feasibility
  • Aggregating data points for technical feasibility
  • Strengthening data protection and beneficiary privacy. If data well anonymize and minimize, then no problem to be aggregated and run through AI for analysis
  • Discussing “ownership” and sharing of data. Serious discussion needs to happen about who has the authority to “own” data, who can responsibly safeguard data (government? NGO? UN? private sector? All but must be compartimentalized? No one?)