How search works at Lanfrica

TLDR: use the search bar for a broad, relevance-based search, and use filters when you need to narrow your results to specific, precise criteria.

This document aims to provide a high-level understanding of how the search feature on the Lanfrica discovery platform works. Understanding this will help you use the search more effectively to find the information you are looking for.

The Lanfrica discovery platform (or simply Lanfrica) is the open-access platform that allows users to explore the landscape and trajectory of AI in Africa, across various domains. Lanfrica organizes Africa’s vast AI resources - datasets, papers, models, and more - under one connected hub.

Simply put, there are two primary ways to search on the platform: using the search bar and using filters.

The Search Bar: Relevance Search

When you type a query into the search bar, the platform performs what we call a relevance search. This means the system finds records on Lanfrica that are relevant to your search terms and orders them based on that relevance. The most relevant results will appear at the top.

For example, if you search for "Ibibio machine translation dataset", the first few results will be the most relevant matches — covering machine translation datasets that contain Ibibio. However, down the list, you will also see results that relate to different parts of your query. For instance, you might see other machine translation datasets that are not for Ibibio, or other Ibibio datasets that are not for machine translation, or linguistic resources on Ibibio.

By default, if you do not specify a different sorting method, the results are always ordered by relevance.

Note: search functionality is currently limited to relevance sorting only. The other sorting options (ascending, descending, and recently linked) are available exclusively for filtering and browsing operations and can not be used while searching.

Filters: Specific Search

If you want to find only the specific results that match all your criteria, you should use filters.

Using our previous example, let's say you want to find only machine translation datasets that contain the Ibibio language. To do this, you would use the filters as follows:

1. Click the Language filter and select Ibibio.

2. Click the Task filter and select Machine Translation.

3. Click the Record Type filter and select Dataset.

The results will then be narrowed down to show only the records that are machine translation datasets containing the Ibibio language.

To give you a richer experience and help you find exactly what you're looking for, we provide a wide range of filters. For example, a researcher searching for climate datasets in Africa can use filters to exclude agricultural works from their results.

Note: Clear the search box before applying filters, as filtering and searching cannot be performed simultaneously at this moment.

Disclaimer: Please be aware that our search results may not always be perfect. We are constantly working to improve their accuracy, and if you spot a result that seems inaccurate, please let us know so we can improve it for everyone.