The starting point: A look at the data in HDX

I hope you had a chance to look through the Humanitarian Data Exchange (HDX) following our launch. Given the large amount of data in the system, I thought it would be useful to provide a brief overview of the main datasets available. This initial collection will serve as a foundation for the new data shared by users as the platform matures.

Defining humanitarian data

As explained in our Terms of Service, HDX defines humanitarian data as: 1) any data that is used to understand the context in which a humanitarian crisis occurs (e.g. baseline data); and 2) any data that is about the people affected by the crisis and their needs, and the response by organizations and individuals seeking to help those people.

Data distinctions

The HDX beta is limited to a dataset repository with basic functionality for visualizations. Data can be shared through individual spreadsheets (i.e. when the data collection is ad-hoc). It can also be shared by connecting local systems through APIs or scraping data. A majority of the initial 1,200 datasets in HDX are shared through an automated process. This allows us to validate the data and present it in different ways, i.e. as an indicator-centric file for all countries and as a country-centric file for all indicators.


The initial release of HDX has mostly national-level data, meaning there is one value that represents the entire country. For example, there were 1,289 reports published by ReliefWeb for Afghanistan in 2013. We will work to add data that is more granular (i.e. broken down by the administrative boundaries within a country) and hope that users will share this type of data.  As a start, HDX includes WFP data that looks at food consumption at national and sub-national levels. More sub-national data will be made available from our pilot countries (Colombia and Kenya) in the coming months.


HDX aggregates data from over 20 trusted sources. This includes data from critical humanitarian partners, such as WFP and UNHCR. We also have important baseline data from the World Bank, UNDP and WHO, among others. And just recently, InterAction shared data from NGO Aid Map. Following our field visits, we are happy to be developing data sharing arrangements with partners such as DANE in Colombia and Concern in Kenya.

The data

There are over 450 tags in the HDX system to help users find relevant data. Enter the following keywords in the search bar to see what is available: population, health, women, and funding. Below is a closer look at a few examples.

Refugee data from UNHCR

UNHCR is sharing data from its population statistics database. Initially this includes three indicators: the total number of refugees, internally displaced persons (IDPs), and asylum seekers from 1960 to 2013. In 2013, UNHCR reported that there are approximately 11.7 million refugees in the world. In the same year, UNHCR protected or assisted about 24 million IDPs. The chart below illustrates these trends from 2000 to 2013. Use keywords ‘UNHCR’ or ‘refugee’ to find datasets related to this topic.

People in need

Food consumption data from WFP

WFP has contributed data from its Food Security Analysis Service. This includes two indicators on households with poor or borderline food consumption. The chart below shows how this indicator varies across 17 countries. Use keyword ‘food’ to obtain more datasets related to these statistics.

Food consumption in selected countries
UN World Food Programme. Food Security Analysis Service / VAM

Crisis statistics from CRED

HDX has six indicators from the Centre for Research on the Epidemiology of Disasters (CRED). These include the number of people killed in disasters, affected by disasters, injured by disasters, and made homeless by disasters, as well as the total number of disasters and the total cost of damage caused. The chart below represents the global trend of disasters from 2000 to 2013. To find more datasets related to crisis statistics, use keyword ‘disaster’.

Total Number of Disasters
Centre for Research on the Epidemiology of Disasters

Please take a closer look at the data in HDX, add new data, and send feedback to And special thanks to XY Feng at Frog Design for the visual explorations included above. We will work with Frog in the coming months to build out the data visualization and analytic functionality of the HDX platform.

We are just getting started – join us!