In April, I wrote about how standards can make humanitarian data easier. Since then, the Humanitarian Exchange Language (HXL) community has made good progress moving those standards from abstract ideas to concrete specifications. First, however, we had to think about where humanitarian standards would have the most impact and bring in the most partners.
For HXL, we are focused on operational data—data that helps us understand what’s happening on the ground during a humanitarian crisis. After consulting with many stakeholders, we determined that the most common path for sharing humanitarian data could be abstracted into three stages, or “miles”:
- First mile: Raw data created in the field (e.g. through surveys, informant interviews or direct observation) goes to an initial collector, such as a cluster lead, or individual NGO, donor or UN agency, to create a local operational picture of a specific geographical location or an area of work (e.g. Health Cluster activities or IDPs in a camp). Multi-cluster needs assessments would also fall into this category.
- Second mile: Collected data from the local actor goes to an aggregator, such as OCHA, a cluster lead or the Government, to combine with other data to form a crisis-wide common operational picture.
- Last mile: Processed data from the local and crisis-wide operational picture is made available to many or all interested humanitarian stakeholders (including media, donors, international headquarters).
Stages in the operational-data path
This idealized view of data exchange does not always apply, of course. But it’s sufficiently detailed to help us choose an initial focus for the HXL standard. The diagram above includes dotted lines to show that data exchange can also happen in other ways (e.g. from NGO to NGO, agency to agency, or directly from a local actor to the world).
Which mile for HXL?
As this is a new standards initiative that needs to show impact to gain momentum, we realized that HXL could target only one of these “miles” fully in 2014. One of our focus data types is 3W (“who is doing what where”) data, describing the activities taking place during a crisis response. Nearly every organization operating in a humanitarian response generates 3W data. Our other focus data type, Humanitarian Profile, describes who is affected by a crisis and how. Not all actors produce humanitarian profile data, but everyone can benefit from consuming it.
To determine where standards would bring the most benefit to these data types, we considered two variables:
1. How big is the potential impact of standards adoption? (The more participants, the bigger the benefit.)
2. What is the technical capacity of the data provider? (The greater the technical capacity, the more likely that providers can adopt standards successfully.)
We determined that these two variables move in opposite directions during the operational-data-sharing process, as shown in the following table:
|Phase||Potential impact||Technical capacity|
The diverse and challenging technical environment (e.g. lack of connectivity, analog vs, digital data collection methods) for organizations collecting first-mile data would make any standards initiative extremely risky. The relatively low potential impact of standards for the last mile means that the effort might not be justified. As a result, standardizing for the second mile was the clear starting point for the HXL community.
Preliminary sketch of the HXL user
Restricting HXL to the second mile in 2014 has allowed us to make some assumptions about the early adopters we are targeting, primarily information managers within OCHA and the cluster and agency leads:
- Information gathering: HXL’s target users have existing relationships with the people and organizations who create the raw operational data in the field, and they already have processes in place for collecting, cleaning and aggregating that data.
- Information management: HXL’s target users already have conventions for organizing and labelling information that are effective within their scope of work. For example, the Education Cluster will often focus on identifying beneficiaries by age and gender. As far as possible, HXL will avoid asking providers to change those conventions.
- Information sharing: HXL’s target users are already accustomed to sharing the type of information targeted by HXL (3W and Humanitarian Profile) to allow OCHA, donors, researchers and their headquarters to create a larger common operational picture. HXL will refine these existing data-sharing processes rather than attempt to define new ones.
- Technical capacity: HXL’s target users all have the technical capacity to create and maintain spreadsheets. Some HXL users may have advanced technical systems, such as data repositories. Many HXL users will not have the technical capacity or available resources to deal with advanced data-exchange formats such as XML, RDF or JSON.
My next blog posting will describe how the HXL community has designed its first draft standards based on these assumptions. In the meantime, I invite you to follow the HXL community discussion by signing up at email@example.com. To find out more about our HXL work, please visit http://hxlstandard.org/