Translation suggestions: Topic-based & Community-defined lists

Group:	Wikimedia Language and Product Localization
Start:	July 2024
Team members:	Runa Bhattacharjee; Niklas Laxström; Pau Giner; Purity Waigi; Sudhanshu Gautam; Abijeet Patrol; Eduardo Medina; Huei Tan; Kartik Mistry; Stephane Bisson; Nik Gkountas; Santhosh Thottingal; Uzoma Ozurumba; Krishna Chaitanya Velaga; Fiona Romeo; ;

Home

Community space

FAQ

Status updates

The Translation suggestions: Topic-based & Community-defined lists is a proof of concept project to validate or invalidate two hypothesis from the WMF 2024-25 Annual Plan Work (WE2.1.2 and WE2.1.4) that:

"If we build a proof-of-concept providing translation suggestions that are based on user-selected topic areas, we will be set up to successfully test whether translators will find more opportunities to translate in their areas of interest and contribute more compared to the generic suggestions currently available."

Welcome to this test project!

This feature developed by the Language and Product Localisation team will help organisers identify and add relevant content based on high-impact topics to their Wikipedia. This proof of concept project will:

Improve the Content Translation tool's article suggestion list support.
Help contributors and campaign organisers discover possible knowledge gaps and find relevant topics in the article suggestion list support as actionable tasks for contributors to work on.
Measure the above approach's impact on content contributions.

"If we developed a proof of concept that adds translation tasks sourced from WikiProjects and other list-building initiatives, and present them as suggestions within the CX mobile workflow, then more editors would discover and translate articles focused on topical gaps. By introducing an option that allows editors to select translation suggestions based on topical lists, we would test whether this approach increases the content coverage in our projects."

Background

In alignment with the Wikimedia Foundation Annual planned objectives and key results under the Equity Goal of supporting the technology needs of the Wikimedia movement:

"Communities are supported to effectively close knowledge gaps through tools and support systems that are easier to access, adapt and improve, ensuring increased growth in trustworthy encyclopedic content."

The Language and Product Localisation team, partnering with the Community Growth team identified the knowledge gaps and content growth as the area to work on. They will explore solving the problem of campaign organizers being heavily burdened with finding actionable tasks for contributors in relevant topics to work on during and after campaigns, and contributors having difficulty in self-discovering workflows and tasks of interest rather than random and generic suggestions of translation tasks.

Rationale for Community-defined translation lists

Organizers coordinate communities' participation around content growth in different focus areas, which can initiate bridges between content gaps or improve existing content in our project platform.

This form of collaboration amongst organizers and contributors, according to the Wikimedia movement organizer study 2019, has the following challenges:
- Editors have difficulty in finding the right content to work on based on chosen focus topics.
- Organizers have difficulty identifying existing contributors in the community with similar interests to participate

Some organizers have devised different tactics and hacky ways to determine their area of focus and build lists of content to work on off-wiki. Still, discoverability, finding the right content to work on and inviting the right existing contributors remain challenges that have to be tackled to help close the content gap in projects.

According to Movement Strategy/Recommendations/Identify Topics for Impact:

One of the 10 recommendations states that we should implement initiatives and prioritize resources to fill content gaps on topics which may have more impact through interventions such as community initiatives, outreach, stipends, grants and other funding, partnerships, and employing methodologies supported by technology, such as artificial intelligence and machine learning.

In the past, the Growth product teams have focused on developing features that improve newcomers' individual editing experience. The Campaigns team's work on the Events center is the closest to working on group collaboration. Therefore, based on the above studies, which have uncovered the challenges of campaign organizers and priorities on resources to fill the content gap, the Language and Product Localisation team will explore interventions for group collaboration to communicate shared priorities to organizers and contributors through the Content Translation tool workflow.

Goals and Objectives

The Language and Product Localisation team will be working on making campaign or topical worklists visible in existing editorial workflows and helping contributors discover them through the improved customized translation suggestions in the Content Translation tool mobile interface where translators land. We will achieve the following:

Increase the discoverability of topics giving editors more options of relevant areas they could contribute to through translation suggestions.
Increase the discoverability of topical gaps already being curated by organizers and communities within events & campaigns.

If we successfully increase the discoverability of these content gaps through improved translation suggestions that contributors can customize. In that case, more editors will pay attention to these worklists and topical content gaps.

Increase quality content contributed using the Content translation tool. If this approach works, we will have proven that topical or campaign worklists are great resources for editors in choosing high impact content to translate.

2.1.2 hypotheses will be validated through the following metrics.

Primary metric:

Number of Quality Articles created from the Content Translation mobile tool originating from topic-selection.

Secondary metric:

Number of pageviews/readership of newly added articles created from translations.

Guardrail metrics:

Deletion rate/ ratio

2.1.4 hypotheses will be validated through the following metrics.

Primary metric:

Number of Quality Articles created from the Content Translation mobile tool originating from community-defined lists.

Secondary metric:

Number of pageviews/readership of newly added articles created from translations.

Guardrail metrics:

Deletion rate/ ratio
Editor signups/ registrations on Campaigns/ Events/ Wikiprojects

Proposed approach

The content translation tool provides article suggestions in the translation dashboard for translators to discover possible knowledge gaps and find relevant topics in those areas to contribute to. We will consider expanding this existing support with additional filtering mechanisms that will provide filtering options to communicate the filter applied and users' customisation options to adjust their suggestion list according to their preferences. The filtering menu will have options like:

The default setting for users with the suggestions derived from:
- For You, based on a user's previous editing patterns
- Popular, based on frequently viewed articles
- Topic level + categories, based on available topic models.

Community lists, based on curated lists of articles needing translation during events/ campaigns. This would also include surfacing content from specific campaigns like Wikiproject campaigns.

Translation Suggestions flow for Community Lists

Multiple selection, based on suggestions from multiple topics. For example, selecting Africa and Art will result in African artists.

Translation Suggestions flow for Multiple Topic selection

Others which can be incorporated in the future based on lessons learned.
- Search using article name or wikidata item to get similar suggestions.
- Selection of topics from a list of Countries or Regions
- Nearby Topics, based on proximity to the user's location using the nearby APIs

Timeline

We target to complete the MVP version in Q1 and begin testing with translation events/campaigns from October onwards.

2024								2025
Q3	Q1				Q2			Q3
Jun	July		Aug	Sept	Oct	Nov	Dec	Jan	Feb	Mar
Ideation and Scoping		Technical Explorations
	Design Explorations & Useability Tests
			Design Iterations
			Product Development
					MVP Testing & Launch
							Feedback, Evaluation & Reporting