Wikimedia Research/Projects

From mediawiki.org

Current[edit]

AI as a service[edit]

Revision scoring as a service

A public, query-able APIs of machine classified scores of Wikipedia revisions

lead:
Aaron Halfaker
team:
Research and Data
collaborators:
Amir Sarabadani, He7d3r, Yuvipanda
home:
m:Research:Revision scoring as a service
other teams involved:
Grantmaking, Ops


Automated classification of edit quality

Build edit reviewing and curation capacity with machine classifiers

lead:
Aaron Halfaker
team:
Research and Data
collaborators:
Amir Sarabadani, He7d3r, Yuvipanda
home:
m:Research:Automated classification of edit quality
other teams involved:
Grantmaking, Ops


Automated classification of article quality

Measure the quality of articles. Direct curation work.

lead:
Aaron Halfaker
team:
Research and Data
collaborators:
User:Nettrom
home:
m:Research:Automated classification of article quality
other teams involved:
Ops


Automated classification of edit types

Measure work types of roles.

lead:
Aaron Halfaker
team:
Research and Data
collaborators:
User:Diyiy, User:DarTar
home:
m:Research:Automated classification of edit types
other teams involved:
Ops


Discussion modeling[edit]

Detox (Modeling Talk Page Abuse)

Understand and model editor interaction via discussion pages and their impact on culture and contributor retention.

lead:
Ellery Wulczyn
team:
Research and Data
collaborators:
CJ Adams (Google Ideas), Lucas Dixon (Google Ideas), Patrick Earley (WMF), Haitham Shammaa (WMF), Dario Taraborelli (WMF), Nithum Thain (Simon Fraser University), Camille Francois (Google Ideas)
home:
m:R:Detox
other teams involved:
CE


Reader research[edit]

Characterizing Wikipedia reader behavior

Identify and characterize different segments in the Wikipedia reader population, by combining qualitative surveys with behavioral analysis based on webrequest data

lead:
Leila Zia
team:
Research and Data
collaborators:
Ellery Wulczyn, Dario Taraborelli, Jonathan Morgan, Robert West (Stanford), Jure Leskovec (Stanford)
home:
m:Research:Characterizing Wikipedia Reader Behaviour
other teams involved:
Reading, Analytics


Contributor value research[edit]

Who adds value to Wikipedia

Analyze historical data to quantify value added to Wikimedia projects

lead:
Aaron Halfaker
team:
Research and Data
home:
m:Research:Measuring_value-added
other teams involved:
Analytics


Contributor merit experimentation

Design experiments to study how indicators of contributor merit/value added affects their motivation (WikiCredit)

lead:
Aaron Halfaker
team:
Research and Data
home:
m:Research:WikiCredit
other teams involved:
Analytics, Editing


Increasing content coverage[edit]

Article recommendation experiments

Identify gaps in content and design recommendation systems to fill them

lead:
Leila Zia
team:
Research and Data
collaborators:
Ellery Wulczyn, Robert West (Stanford), Jure Leskovec (Stanford)
home:
m:Research:Increasing article coverage
other teams involved:
Editing (CX), Ops (Labs)


Improving access to content[edit]

Improving access to content

Identify bottlenecks in navigation and design recommendation systems to optimize the reader experience

lead:
Leila Zia
team:
Research and Data
collaborators:
Robert West (Stanford), Ashwin Pradeep Paranjape (Stanford), Jure Leskovec (Stanford)
home:
m:Research:Improving link coverage
other teams involved:
Reading, Analytics


Research management and collaborations[edit]

Research consulting

Provide ad hoc research support to other teams at WMF

lead:
Dario Taraborelli
team:
Research and Data
collaborators:
All team members
other teams involved:
WMF at large


Research collaborations

Handle external research collaborations, academic/industry outreach effort

lead:
Dario Taraborelli
team:
Research and Data
collaborators:
All team members
other teams involved:
Legal


Research management

Team management-related tasks

lead:
Dario Taraborelli
team:
Research and Data
collaborators:
Abbey Ripstra


Data releases and API prototypes[edit]

Data releases

Release curated datasets; prototype, test, and evaluate new data-intensive APIs

lead:
Dario Taraborelli
team:
Research and Data
collaborators:
All team members
other teams involved:
Legal, Ops (Labs), Analytics


Wikistats maintenance[edit]

Wikistats

Maintain through the end of the calendar year the legacy reporting infrastructure for Wikimedia metrics

lead:
Erik Zachte
team:
Research and Data
other teams involved:
Ops, Analytics


Generative Design Research[edit]

Persona research

Research our users to inform a solid and inclusive set of robust personas. Currently we have "pragmatic personas" which are our best guess at personas reflecting our users from what we know - and being honest about what we do not know. This persona research will allow us to better gain knowledge about the user we do not currently know enough about. File:WMF pragmatic personas product.pdf

lead:
Abbey Ripstra
team:
Design Research
other teams involved:
Communications, Audience teams, Community Engagement


Support product teams with generative design research

Collaborate with product teams to implement generative design research (for example contextual inquiry, exploratory interviews, design ethnography, card sorts, surveys) within product development.

lead:
Abbey Ripstra
team:
Design Research
other teams involved:
Communications, Audience teams, Community Engagement


Evaluative Design Research[edit]

Support product teams with evaluative design research

Collaborate with product teams to implement evaluative design research (for example usability testing, concept evaluation, rapid iterative testing) within product development.

lead:
Abbey Ripstra
team:
Design Research
other teams involved:
Communications, Audience teams, Community Engagement


Supporting editor workflows[edit]

New editor curation tools experience interviews

Interview active editors who participate in community processes to inform the development of a workflows feature for Flow.

lead:
Jonathan Morgan
team:
Design Research
home:
meta:Research:New editor curation tools experience interviews September_2015
other teams involved:
Editing (Collaboration)


Flow workflow interviews

Interview active editors who participate in community processes to inform the development of a workflows feature for Flow.

lead:
Jonathan Morgan
team:
Design Research
home:
Flow/Community_process_workflow_interviews_(June_2015)
other teams involved:
Editing (Collaboration)


Completed[edit]

VisualEditor experimentation[edit]

Visual Editor experimentation

Design, implement, and communicate a robust experimental strategy for the Visual Editor rollout

lead:
Aaron Halfaker
team:
Research and Data
home:
m:Research:VisualEditor's effect on newly registered editors/May 2015 study
other teams involved:
Editing (VisualEditor), Analytics


Fundraising research[edit]

Fundraising research

Experimental design and optimization for the online fundraiser.

lead:
Ellery Wulczyn
team:
Research and Data
other teams involved:
Fundraising


Planned[edit]

Contributor routing[edit]

Newcomer motivation prediction

Predict the new contributors' motivation to route them to the appropriate onboarding/socialization flow

lead:
Aaron Halfaker
team:
Research and Data


Editor-role based task routing

Identify roles needed for each article and design an automated recruitment workflow

lead:
Aaron Halfaker
team:
Research and Data


Newcomer onboarding experimentation

Experiment with different onboarding/socialization flows for newbies

lead:
Aaron Halfaker
team:
Research and Data


On hold[edit]

Data releases and API prototypes[edit]

Geo-aggregation of traffic and contribution data

Design an anonymization/aggregation strategy and release geodumps of traffic and contribution data

lead:
Dario Taraborelli
team:
Research and Data
collaborators:
Reid Priedhorsky (Los Alamos); Mark Graham (Oxford), Shilad Sen (Macalester)


Evaluative Design Research[edit]

(REFLEX)

Benchmarking tasks and workflows that are important for each persona to be able to do. To evaluate the user's experience in completing (or not) those tasks and workflows in production.

lead:
Abbey Ripstra
team:
Design Research
other teams involved:
Audience teams


Generative Design Research[edit]

Deep Dive (Design ethnography / Contextual inquiry)

Gain knowledge about far flung users (informing personas) of wiki projects, their contexts (around knowledge, technology, access) and what they need to be successful in contributing to and gathering knowledge from Wikimedia projects

lead:
Abbey Ripstra
team:
Design Research
other teams involved:
Communications, Audience teams, Community Engagement