Aid
Data
A Research Lab at William & Mary

The AidData SDG widget visualizes aid flow estimates across the Sustainable Development Goals based on a mapping of AidData activity and purpose codes to SDG targets.

The data source for this information is the Financing to the SDGs Dataset Version 1.0.

The visualizations are divided by tab, allowing users to facet and filter the data according to donors, recipients, years, and SDGs in a few different ways.

1) SDG Funding by Donor affords an intuitive view of relative funding across SDGs for a given donor, recipient, and/or year(s). A dynamic tree chart responds to these user selections by displaying SDGs as tiles, where each SDG tile is scaled to the proportional funding amount.


2) Top/Bottom Donors visualizes the top or bottom donors that provide funding to a particular SDG for given recipients and years utilizing a bar graph visualization.


Methodology in Brief:

Tracking and analyzing funding to the Sustainable Development Goals (SDGs) will be central to measuring progress, crowding in resources to priority areas, and helping decision-makers make more informed choices. Unfortunately, currently available data do not capture sufficient information on the distribution of financing for the SDGs. The AidData Sustainable Development Goals Estimates attempt to fill this gap by providing project-level estimates of contributions to the SDGs and their associated targets, letting us see where development financing is targeted and allowing comparisons among different SDG goals.

This methodology is based on an analysis of development project descriptions and builds on an existing activity coding schema developed at AidData. Activity codes are based on the OECD's Creditor Reporting System (CRS) sector and purpose codes, but go one step deeper, providing a more disaggregated breakdown of development activities that are relevant to each CRS code. Student coders had previously activity coded 58% of projects in AidData's core research release.

The methodology involves three main steps:

  1. Creating a mapping between activity codes and SDG targets;
  2. Splitting an aid project across designated activities; and
  3. Splitting activity amounts across SDG targets, as an activity may be linked to multiple targets.

Where activity codes are unavailable, we use purpose codes, which are not as granular as activity codes, and generate estimates based on a naïve diffuse assumption about what activities were involved in a project with a given purpose code.

While we recognize that components of an aid project may receive differing levels of resources in practice, the available project documentation seldom contains this level of granularity on how money is distributed at the sub-project level. Therefore, we split dollar values equitably across activities as the best available option in the absence of perfect information. Having split the dollar value of a project across unique activities, we then distribute those activity-dollar amounts evenly across the SDGs targets associated with the activity. From these calculations, we can sum target-level estimates up to the goal level.

For more detailed information on the statistical methodology, please refer to our methodology paper Estimating Baseline Aid to the Sustainable Development Goals. AidData is also currently working on a new iteration of this methodology, which will code projects to the SDGs directly rather than through a crosswalk.