User:KeerthanaS

From MediaWiki.org
Jump to: navigation, search
Keerthana User.jpg

I am currently in the sophomore year of my undergraduate program. My major is Engineering Physics and I study in Indian Institute of Technology, Madras, one of the premier institutes in India for Engineering. I got interested in coding essentially because of the abundance of the open and free resources available for developing your skills in that and how easy it is to create technologies and products only with the knowledge of coding. I publish some tutorials on the technologies I am working on in my blog.

Summer of Code 2017

  • The Phabricator task for the project can be found here here.
  • The proposal for the same can be found here.

GSoC 2017 Project : Automatic editing suggestions and feedbacks for articles in Wiki Ed Dashboard

Synopsis

The Wiki Ed Dashboard / Programs & Events Dashboard is a Ruby on Rails + Javascript application that helps people organize groups of newcomers to contribute to Wikipedia.

Aim of the Project

This project aims to use ORES (Objective Revision Evaluation Service) to provide

  • To add a feature to give some automatic feedback on the articles being edited by the dashboard users using the ORES based data on the articles.
  • Using the same data to build a feature that can help editors find an article to which they can contribute the best to based on their previous Wikipedia editing experience.

Mentors

Sage Ross, Jonathan Morgan.

ORES data

The ORES data as of now provides predicted article classes and also has a data on the features used for the prediction.

MyArticles component

The newly added MyArticles component is a great place to start with for providing feedback on some revisions. It is an accumulation of all the article an user is creating, editing and reviewing. For the first part, the idea is to use the features provided by the ORES API to present to the users some suggestions on what areas to improve their articles based on the features list (For example some feedback on citations and structure). For this the initial approach is to use our judgement to give some advice on the appropriate values on these features. Then the ORES predicted rating based on the latest revision will be used to give some editing suggestions.

Feedback on the suggestions

The automatic feedback feature can be definitely useful for the newcomers to Wikipedia editing and presenting these information in a way they can easily find it and use it is gonna be a major focus of this project. Getting some feedback on these suggestions from several people should be also be a focus on improving these feedback messages. To enable proper feedback focus should also be on presenting the feedback to instructors and other Dashboard users who are interested in the development of the article who might be able to give useful feedback to the suggestions given. How it is going to be implemented in UI is not yet decided.

Using the feedback

The feedback received is going to be a major tool in building the suggestion messages presented and can be a mechanism for the continuous improvement of the suggestion messages even after the project. Though a generic feedback system for the dashboard exists. The current feedback should be presented in the admin dashboard in the context of the article, its ORES data and the suggestions provided in order to use the feedback to improve the suggestion messages.

Available Articles component

The Available Articles section is a place to start with for suggestions based on article rating and improving on it with some ORES features based suggestions. This is so that students can pick articles based on what kind of revisions they want to make. The same kind of suggestion as in the MyArticles components can be applied here except it can use the imported article rating. While for the articles being edited and sandbox articles the talk page is going to be the major source for instructors to give direct feedback on the article for Available articles there can be an option for instructors to add some of their own suggestions in the suggestions box.

Schedule

Timeline Task Remarks Work report Status
Week 1 (Due May 18) Community Bonding period

Introduced myself to the community through mailing list. Joined Zulip. Discussed with mentors about participation during GSoC. Learning new technologies for the project.

First Blog post

Yes check.svg Done

Week 2 (Due May 25)

Community Bonding period

Updated deliverables. Did some research on the dashboard and finalized the goals of the project.

Week 2 - Blog post

Yes check.svg Done

Week 3 (Due June 1)

Design and add buttons in the MyArticles component that would display the feedback

The feedback buttons will be a way for users to say how each and every aspect of the suggestion was useful and would be designed in a more sophisticated manner in the future.

Blog post

Yes check.svg Done

Week 4 (Due June 8)

Add feedback based on ORES features

The features are retrieved and processed in the rails backend to generate feedback. They are not stored in the backend as of now. They are retrieved in the frontend through Redux. The features based feedback should be improved.

Blog post

Yes check.svg Done

Week 5 (Due June 15)

Add feedback based on ORES rating

The ORES predicted rating is retrieved from the ORES API and was displayed in the frontend

Blog post

Yes check.svg Done

Week 6 (Due June 22)

Improve Feedback messages

The feedback messages are improved and design for displaying it improved

Blog post

Yes check.svg Done

Week 7 (Due June 29)

Extend the feedback feature to Sandbox articles and add feedback button to Assigned Articles component.

Extended the feedback feature to sandbox articles. Improved the suggestions to give the users a hint on the Wikipedia Assessment Scale. Added a way to receive suggestions from the users within the modal.

Blog post

Yes check.svg Done

Week 8 (Due July 6)

Improve the feedback for further testing. Ideation to improve automatic feedback.

An open Excel Sheet to get some resources and all hands on improving the automatic feedback

Blog post

Yes check.svg Done

Week 9 & 10(Due July 20)

Add Custom Feedback Feature

Users can add a feedback to an article in addition to the existing automatic feedback

Blog post

Yes check.svg Done

Week 11 & 12(Due August 3)

Improve the custom feedback feature

Added user reference to feedback Added delete functionality Finalized project deliverables with mentor

Yes check.svg Done

Week 13 (Due August 10)

Improve the Automatic Feedback

August 10 to August 21

Identify areas of improvement in ORES. Write documentation

Final Week

Submit Code and Evaluations