HDX is a Year Old: Our stats and a look ahead

When we launched the Humanitarian Data Exchange (HDX) one year ago at the Open Knowledge Festival in Berlin, our goal was to make humanitarian data easy to find and use for analysis. A year later, we’ve made a huge amount of progress towards that goal, but as the United Nations is fond of saying “more work remains.”

Our first-year focus was getting data into HDX — without useful data, we knew that no number of features and or amount of design would make the site worthwhile. We began by working with partners in two locations, Colombia and Mali. We established a Data Lab in Nairobi to help uncover relevant data in East Africa. We developed crisis pages for the Ebola Outbreak in West Africa and for the Nepal Earthquake. We also created branded organization pages, initially to showcase the World Food Programme’s food data but subsequently for many more partners.

A year of stats

We track a number of statistics to help understand user behavior and to ensure we are having an impact. Below is a close look at few key indicators:

The number of public datasets on HDX – We launched the platform with 1,174 datasets. This collection has grown by 787 to reach the current 1,961 datasets on the site. HDX holds 54 datasets related to the Ebola Outbreak and 81 related to the Nepal Earthquake. UNOSAT alone has contributed 160 datasets, mostly geospatial data from its satellite imagery, such as a damage assessment of the Sana’s airport in Yemen.
Number of Public Datasets on HDX

The number of registered organizations on HDX – We launched the site with 16 organizations and now have 180, 60 percent of which are actively sharing data. These organizations range from large UN agencies (e.g., UNHCR) to governments (e.g. the U.S. State Department and the Kenya Open Data Initiative), to a variety of NGOs (e.g., Humanitarian OpenStreetMap and Interaction) and research institutions (e.g., Columbia and Oxford universities).
Number of Registered Organizations on HDX

The number of users and the number of registered users – We have had over 80,000 users visit the site since last July, with peaks during the height of the Ebola crisis and the more recent Nepal Earthquake. We have also seen a steady increase in the number of people who have created an account on HDX. Registered users are able to access more features like sharing data (through an organization) or following data for updates. We just introduced a new user onboarding process. If you haven’t yet registered, try it!
Number of Users

Number of Registered Users

The proportion of site users who downloaded data – On average, a quarter of all visitors to the site download data. We are most encouraged by this statistic as it shows that the data being shared through HDX is relevant to our users, something we care deeply about. The dataset on the Number of Ebola Cases and Deaths in Affected Countries has been downloaded almost 3,000 times. For the Nepal Earthquake, the top dataset to be downloaded is the Nepal: Official figures for casualties and damage.
Proportion of Site Users who Downloaded Data

A look at what’s to come

While we are encouraged by the numbers above, we want to do more for our existing users, and reach out to more people who can benefit from humanitarian data and data services. Here is what we have planned for the year ahead:

  • We are releasing a new feature for previewing geospatial data onto a map for immediate insight. Take a look at an example here.
  • We are migrating the operational datasets stored in OCHA’s Humanitarian Response site to HDX to make it easier to find humanitarian data.
  • We will be improving search, and making it easier to find and filter data that is being actively maintained.
  • All organizations in HDX will be able to showcase their data through branded pages that include their own logos and color schemes.
  • We will add functionality for HXL-tagged data to be processed and visualized through HDX.

Our team is motivated to take on these new challenges. We are grateful to the many users who have provided feedback and helped us stay user-driven, and we owe a special thanks to Aidan McGuire at Scraperwiki and Rufus Pollock at Open Knowledge for keeping us on solid technical footing.

Send us feedback at hdx.feedback@gmail.com or reach us on Twitter at @humdata.