Geographic Unit Boundaries and Definitions

Managing Geographic Units for the FEWS NET Data Warehouse (FDW) includes the following tasks:

Introducing new Geographic Unit boundaries and definitions
Editing existing boundaries
Editing existing relationships defined between Geographic Unit boundaries
Associating Geographic Unit boundaries with new and/or existing data in FDW

This guidance provides an overview of methods for preparing Geographic Units and relationships for upload to FDW. It is intended for Hub staff members and partners who add, delete, update or substantially change new or existing:

FDW spatial boundaries and definitions
FDW statistical data
Relationships between FDW data and FDW spatial reporting boundaries

Although this guidance references spatial Geographic Units and statistical data, in most cases the information is equally applicable to Geographic Units without any true spatial character, and numerical data of any kind.

When to upload Geographic Unit boundaries and relationships

Geographic Unit boundaries and relationships need to be uploaded to FDW in the following cases:

The Geographic Unit boundaries do not exist in FDW.
The Geographic Unit boundaries do not cover the period of time reflected by new data.
The Geographic Unit boundaries have an error in some feature(s) of their names, coding, hierarchy definitions, or relationships over time.

Checking existing spatial data

For most of the new data that will be added to the FDW, such as price data or annual crop data updates, the FNIDs, relationship tables, and less often, the Geographic Unit Set Version will have already been set up in FDW. In that case, you must confirm whether the new data to be uploaded is accurately coded with FNIDs which match what is already in FDW.

You can find the FNIDs already defined in the FDW for a country’s subnational data by searching the https://fdw.fews.net/data-explorer/spatial and using the API link at the bottom of the page.

If the data to be uploaded has FNIDs that perfectly match those in FDW, you can proceed with uploading the new data to FDW.

If there is a mis-match between the new data and the Geographic Units, FNIDs or Geographic Unit relationships already found in the FDW, you must:

Determine the best way to adjust the new data or what is in FDW so they will match.
Coordinate any changes to the Geographic Units and data already in FDW with the Hub Data Team.

Overview of preparing Geographic Units and relationships

In general, the FDW supports two ways of uploading Geographic Units:

Spatial representation (polygon boundary files): When available, upload a complete set of Geographic Unit boundary files shows the polygon boundaries that were in use during each year within the desired period of data coverage.
Tabular representation (spatial boundary definitions only): If a complete historical set of spatial boundary files is not available for a country, you can upload Geographic Unit data to FDW using specified geographic boundary definitions in a spreadsheet format.

When providing boundary definitions only, it is not possible to visualize their data on a map or conduct spatial analyses of the stored data.

Preparing spatial representation

The Geographic Unit boundary files must show the polygon boundaries that were in use during each year within the desired period of data coverage.

Several Geographic Unit Set Versions may be needed to cover the entire period, depending upon the frequency with which the boundaries have changed over time.

Each file for a Geographic Unit Set Version may have several boundary layers, depending upon the number of levels in a country’s administrative hierarchy (e.g. national-level boundary, 1^st-level administrative unit boundaries, 2^nd-level administrative unit boundaries, etc.).

Prepare and submit the following to the Hub Data Team:

A complete set of Geographic Unit Set Version boundary filesin a spatial format, such as an ESRI Shapefile.
An FDW relationship table which describes the evolution of Geographic Unit boundaries in a country over time, in the prescribed format.

The Hub Data Team will set up FDW to manage the new data.

For more information, see Geographic Unit Set Versions: Administrative and Crop Production and Geographic Unit Relationships.

Preparing tabular representation

If you don’t have a complete set of Geographic Unit Set Version boundary files (i.e., they do not cover the entire period of time described by the data, inclusive of all boundary hierarchical levels) you must develop the following:

A boundary definitions file in CSV or XLSX format.
A relationship table.

To complete both the boundary definitions file and the relationship table, someone must identify the individual boundary Geographic Units (e.g. administrative entities, crop-reporting units, etc.) that comprised the country’s entire evolution of Geographic Units over the desired period of time.

Answer the following questions when preparing Geographic Units:

How many total Geographic Units existed in the country, in each year of the desired period of time?
What names do/did the Geographic Units have in each year?
Where are/were the Geographic Units in the country’s administrative (or non-administrative) hierarchy?
Over the chosen period of time, which Geographic Units changed their geographic shape or name, and when?

For more information on preparing the boundary definitions file, see Geographic Unit Set Versions: Administrative and Crop Production.

Challenges with boundary changes

There are particular challenges to determining how and how many times the Geographic Unit boundaries have changed in the desired period of data coverage.

Boundary changes can occur for a variety of reasons. The key changes to identify are those where a single Geographic Unit changes its shape. These changes include:

Splits: One Geographic Unit divides into two or more units.
Merges: Several Geographic Units become one.

There are a number of other changes that are significant to note if they happen. For guidance on all the types of boundary changes that the FDW can record, see Geographic Unit Relationships.

If the Geographic Unit’s evolution is known, then time-series analyses can be accurately assessed. For example, changes in yield per unit of area sown or arable land per farmer over time.

Sometimes, however, the identified changes may later prove to be inaccurate, and it becomes necessary to repair an existing set of Geographic Unit relationships as described below.

Repair existing Geographic Unit relationships

In some cases, the best solution will be to delete what exists in the FDW already, and reload an accurate set of Geographic Unit Set Version boundary units and their relationships.

First, however, you must assess the data linked to the existing Geographic Units. If you delete the existing set of Geographic Unit relationships in FDW and re-upload the corrected set, the linked data will be lost.

If reloading is impossible, there are two other options:

Change the existing dataset so it aligns with the units from FDW, or;
In the case of crop data, create a new set of crop_boundaries or definitions that would be used for this type of data only.

Coordinate with the Hub Data Team on the solution you think would be best.