How to link a record on Lanfrica

Follow this step-by-step guide to "link" a record to the Lanfrica platform.

In order to link a record, you need to be signed in. This tutorial assumes you are already signed in to your Lanfrica account. For a guide on creating an account, refer here.


Click the "Link Record" button from either your Dashboard or your "My records" page. If using a mobile device, then it is the "+" icon at the bottom of the page. Clicking it launches the platform for adding/linking your record. It is meant to be a simple, quick process.

Main URL

The first metadata required of the record you are adding is the main URL that points to the record. When you input this, we will first check whether Lanfrica already has the record, before moving to the next steps.

In addition to the main URL, you can add more URLs that describe the record.

Note: you will be able to also link un-digitized records, with no URL. More on that in the future.

2. Input Record Metadata

Here, you put the following information. Feel free to copy and paste the information from where the record is hosted:

  1. Title: Provide a title for your record.

  2. Short Description (or abstract): Give a brief overview or summary of the record. For research papers, you can paste the abstract here.

  3. Click “Continue” to move on.

A clear and detailed description helps others understand the content and relevance of your record.

3. Organizing the record

A significant value of Lanfrica is that it has structured, organized information about the record — this detailed information enables analysis and better understanding of the record.

Type of Resource

This field identifies one or more categories that best describe a given record. Although a record may span multiple types, select the category or categories that most accurately represent its primary nature. Available record types include:

Record Type
What it means

Dataset

describes records that are machine learning or artificial intelligence datasets in raw or processed forms.

Paper

describes records that are papers. This includes research papers and books.

Software

describes records that are software packages.

Project

describes projects in their different stages (from ideation to completion). For example, GitHub repositories typically fit this category.

Poster

describes records that are posters. This includes research posters. A good example are the Deep Learning Indaba posters.

Policy

refers to policy documents of any kind. Usually we want policy documents that pertain to the African continent.

Language assessment:

this is a rare type that refers to records that are platforms for assessing language skills.

Media

describes media records. These could be media articles, newspapers, etc.

Tools

describes records that are tools, including linguistic tools (like keyboards and dictionaries), as well as other tools (physical or digital).

We aim to accommodate a broad spectrum of record types. If you believe an important category is missing, please let us know at [email protected].

Domain

Choose the domain that is the best category for your record. The full list of our domains and their meanings can be found here. Available domains include:

  1. Language:

  2. Agriculture

  3. Health

  4. Finance

  5. Natural language processing

  6. Climate

Please note that a record may be classified under more than one domain. Also, any record in the Natural Language Processing domain also falls under Language (the reverse is not necessarily true), since NLP records invariably involve one or more languages.

Tip: choosing the right domains for your record ensures that it shows up in the correct searches.

Modality

The Modality field specifies the forms of content contained in the record — particularly important for datasets. It answers the question: What type of data is contained in this record?

Supported modalities:

  • Text

  • Speech

  • Video

  • Image

African Language

This field applies when you have selected Language or Natural Language Processing as a record type. Here, list the language(s) that the record addresses or analyzes. Note that this refers to the subject language(s) under discussion—not necessarily the language in which the record is written.

Example: A research paper on Xhosa–English machine translation, though authored in English, would identify Xhosa (and/or English, if both are studied) in this section.

Note: If the language you want isn’t listed, contact us to [email protected] to add it.

There are two ways to add the language:

  • One-by-One

    • Click the language dropdown.

    • Search for your desired language.

    • Click to select it.

    • Repeat these steps to add additional languages individually.

  • Bulk Addition Sometimes, the number of languages covered by a record are too many to conveniently add one by one. For this scenario, we provide a way to add multiple languages in bulk.

    • Click the Add Multiple Languages icon to open the bulk-entry dialog.

    • Paste your list of ISO codes or language names into the text box.

    • Choose the appropriate delimiter (for example, select Comma if your entries are comma-separated).

    • Click Extract. The system will parse your input, match each entry against our available languages, and populate the dropdown.

    • Review and refine the results (e.g., add any languages that were missed).

NLP Tasks

Specify the natural language processing task(s) relevant to this record (for example, Machine Translation, Sentiment Analysis, Named Entity Recognition, etc.). This only applies to natural language processing resources.

Example: If your record is a text classification dataset in Swahili, you might choose “Text Classification” as the NLP task and “Dataset” as the resource type.

License

Select the license that governs this record. There are two ways to specify the license:

  • Write just the name of the license. This makes sense if it is a well-known license (e.g., CC-BY, MIT, GPL)

  • Write the URL to your license. This allows for custom licenses, or if you prefer, to provide a link.

Tags

Tags are keywords that help other users find your record. Add words that describe what your record is about or what it's used for - the more relevant tags you add, the easier it will be for others to discover your content.


4. Need Help?

If you run into any issues or have questions:

Last updated

Was this helpful?