AfSIS DB • QED

AfSIS DB

The world’s largest public repository of African soil chemistry data.

Including dry and wet chemistry measurements, and their georeferences.

Decades of field and laboratory work, accessible to the world, for free.

Overview

AfSIS DB is a web portal providing an interface for uploading, searching, and downloading soil chemistry data. It provides a lightweight and centralized laboratory information management system in the cloud for soil laboratories affiliated with the AfSIS project, including The World Agroforestry Centre (ICRAF), the Tanzania Soil Information Service (TanSIS), the Ghana Soil Information Service (GhaSIS), Rothamsted Research, and One Acre Fund (OAF).

If you are only interested in downloading datasets of paired dry and wet soil chemistry, please consider the Amazon S3 bucket at

arn:aws:s3:::afsis

Thanks to the gracious support of Amazon and collaboration efforts from all of our partners, we are very happy to announce that this data is now available through the Registry of Open Data at AWS (RODA), to be hosted by Amazon ad infinitum for public dissemination due to its importance as as public good.

For more information about what is available at the RODA S3 bucket and how to use it., please visit:

https://github.com/qedsoftware/afsis-soil-chem-tutorial

The georeferences table stores the latitude and longitudes of all soil sample numbers, allowing multiple labs to know where soil samples came from. Simple map visualizations for these locations are also automatically created.

Spectroscopy readings collected from proprietary hardware are stored their original, raw binary formats in S3. We do this to preserve numerical precision and avoid transcription errors. Upon upload, input validation code is executed to check for possible misconfigurations of the spectroscopy hardware.

AfSIS DB supports multiple kinds of searches across data tables:

SSN Search: Find all records corresponding to a particular SSN (sample serial number). The same SSN may be analyzed by multiple instruments at multiple labs.
Circle Search and Bounding Box Search: Find all records in the database that lie within a particular geographic circle or bounding box.