The AFR IDEAS (Improved Data Extraction and Screening) microlab emerged from the workflow developed by the main CE AFR Lab, which worked to gather and synthesize an unprecedented body of literature about reforestation and forest regeneration around the world. This process relied on both manual extraction of articles from databases, as well as automated scraping of databases using scripts, followed by human analysis to screen for the particular content within each source. Unsurprisingly, this process took a lot of time and effort from the main lab team.

This effort-intensive process of gathering literature and extracting key details (‘metadata’) is common to many literature review efforts, and multiple tools exist to automate and streamline both identification of in-scope literature and recording metadata. One leading open-source tool is ASReview, which was used in a complementary workstream to the AFR Lab carried out by academic collaborators. ASReview has a sophisticated machine learning algorithm to identify sources relevant to a literature review, but was not used for the main AFR Lab because it did not provide a way to store structured metadata.

The AFR Ideas team decided to build this feature to add to ASReview. They created an interface for reviewers to specify structured data to add to sources, for example specific subject, region, scale, and relevance markers, rather than having to add whole text notes. This enables faster review, more consistent metadata extraction, and more streamlined search and analysis of sources. Because this data is stored in the backend along with other metadata for each source, it also has the potential to be utilized by the machine learning model to give it the capability to scan source titles and abstracts and pre-suggest tags. These open-source additions improved ASReview’s functionality for researchers working in the reforestation domain and beyond.

We are happy to report that scientists using the new software have had very nice things to say, such as, “this will open up many possibilities and is a useful feature for many people.” And “This feature is amazing. Thank you for contributing.”

Evaluating

Along with training 60 community field scientists to collect water samples, we will also partner with government and civic organizations to maintain the green infrastructures. Maintaining the new systems of pollution abatement as well as new parks along the edges of the naalas are essential to ensuring their continued effectiveness.