[Objective 9.2.2] Personalized recommendations
Closed, ResolvedPublic
Actions

Assigned To

Authored By

	• DarTar
	Jul 20 2017, 9:43 PM

Description

About

Develop (personalized) recommendations for the editor dashboard tool, in collaboration with the Editing team. This objective may involve exploring new types of recommendations. (consulting)

Background
As part of PP5, a dashboard (or system) will be designed to surface a subset of backlog of work by community. Such dashboard or system can be a place where recommendations for what to do next can be exposed to editors. To this end, objective 2.2 of TP9 was developed to surface specific type of recommendations developed by Research to interested editors via the dashboard/system.

What kind of recommendations can the research immediately surface in such a dashboard?

Article recommendation for creation from scratch (with the option of personalization for the user or not)
Article recommendation for creation via translation (with the option of personalization for the user or not)

What kind of recommendations can become available for testing in the dashboard?**

It would be really helpful if some space in the dashboard is allocated to experimentation, to surface ongoing research that can become permanent part of the dashboard if successful. To this end, we can start testing certain types of recommendation:

Article expansion recommendations: we are very close to be able to start testing with section recommendations, for example. On top of that, we are planning to start experiments on image recommendation (for images to be added to articles), infobox content recommendation, etc. (Pau has already provided some designs for section recommendation)
Hyperlink recommendation (Nirzar and Pau worked on design recommendations for a tool that could surface such recommendations some time ago)
...

Subtasks:

[Leila] Discuss the ideas in more details with Joe
[Joe] To get back to Leila with a decision whether he sees an opportunity for the dashboard to start surfacing recommendations, starting with recommendations on what article to create next
[Leila and Joe] Define next steps based on the the decision above.

People:

@leila
RS

Dependencies:

Contributors/Editing

Estimated quarter(s):

Q1-Q4

Notes:

In 2015, we ran a large experiment through which we showed that personalized recommendations can increase article creation rate in Wikipedia. In the set-up of the experiment, we showed that such recommendations can increase article recommendations by a factor of 3.2. (Documentation of research and results) Through this research, we know that giving the option to editors to personalize recommendations is important, we also know that recommendations work.

Event Timeline

• DarTar created this task.Jul 20 2017, 9:43 PM

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptJul 20 2017, 9:43 PM

leila updated the task description. (Show Details)Aug 30 2017, 11:20 PM

leila added a subscriber: • jmatazzoni.

@jmatazzoni I've updated this task based on the discussions we had today. I added some more details in the description. By September 8, I need to propose the goals for this objective that need to get done in Q2 on our end and for that, I need your input. :)

One question that came out of our meeting today that you mentioned you will think about is:
Do you see a possibility for the design and development of the dashboard to get started on your end, with the goal for the dashboard to surface the types of recommendations we know how to extract, prioritize, and recommend (article recommendation for creation) and add to the dashboard more types of tasks as more types become available through your research (of other community backlogs) or ours (section recommendation for article expansion, etc.)?

Do you think you can get back to me by Monday, September 4, with an answer to the above question? That can help us prioritize the goals associated with this objective on our end for Q2 (the goals for Q2 need to be proposed by September 8)? If this objective is to be met on our end, we need to start the work on it with you in Q2. :) Please let me know if this timeline works for you and/or if you need other information from me.

In T171229#3568591, @leila wrote:

Do you see a possibility for the design and development of the dashboard to get started on your end, with the goal for the dashboard to surface the types of recommendations we know how to extract, prioritize, and recommend (article recommendation for creation) and add to the dashboard more types of tasks as more types become available through your research (of other community backlogs) or ours (section recommendation for article expansion, etc.)?

Yes, I think that surfacing intelligent article-creation (and translation) recommendations would be an excellent candidate for inclusion in a general task-suggestion dashboard, which we will be researching next quarter pursuant to our quarterly goal Goal 5.1.1, and would plan to begin building in Q3. I have to offer the caveat that I've done very little product research on this goal yet, so I'm not in a position to make a commitment.

leila closed this task as Resolved.Jan 15 2019, 10:07 PM

[Objective 9.2.2] Personalized recommendationsClosed, ResolvedPublicActions