Intro

• Data created by regular users of web technology (non-professionals)

• comprises content (e.g. text, audio, image, video) and as well corresponding metadata (user-system interaction, user identifiers, sensor data from smart devices)

• desire to express oneself

• connecting with peers

• achieving a certain level of fame, notoriety, prestige, or influence

• making money (“influencer”) → still within the definition of UGD?

• Publication requirement: work published in some context to a considerable sub-group of users (excludes email and instant messaging)

• Creative effort: some form of added value is needed for creation or adapting exisitng work. Just reposting of clips of tv shows is not considered as UGC

• Creation outside of professional routines and practices: Created by general public. Not produced institutional or commercial market context. Perhaps without expectation of profit or remuneration

• Volume: Scale of data

• Variety: different forms of data

• Velocity: analysis of streaming data

• Veracity: uncertainty of data

• recsys

• influential user detection

• hate speech detection

• fake news detection

• System that recommends interesting items to the user (filter relevant items)

• Created from implicit or explicit user generated data

• Important factors are similarity, diversity, popularity etc,.

• Double-counting: same UGC is sometimes accessible on a variety of sites (duplicate detection)

• Inactive accounts: not all registered users are active (→ MAUs)

• Counting unique users: same user has multiple accounts

• Fake accounts, used e.g. for malicious attacks such as increasing views

• Community bias → statistical bias

• Popularity bias → unfairness

• Copyright: issues with tracking origins of UGC

• Privacy concerns: too much personal (sensible) information accessible publicly