生产运营/行业运营>> Master-Thesis>> 返回
项目作者： Medical-Hippo-1498

项目描述 ：
The Added Value of User-Generated Content in Online Adult Entertainment
高级语言： Python
项目地址: git://github.com/Medical-Hippo-1498/Master-Thesis.git
创建时间： 2021-07-01T11:35:04Z
项目社区：https://github.com/Medical-Hippo-1498/Master-Thesis
开源协议：
下载


Title: The Added Value of User-Generated Content in Online Adult Entertainment

Description:

This GitHub project was created in the context of my Master Thesis for the program of Business Intelligence and Smart Services at Maastricht University.

In this thesis, I fill a gap in the academic literature by deriving the added value of user-generated content (UGC) in the online adult entertainment industry and by investigating the idiosyncrasies of this research context with regard to topic modelling and sentiment analysis using data from an online adult entertainment platform as well as YouTube comment data. Moreover, drawing on the significant predictive power of UGC on the popularity of online adult entertainment videos, I develop a data analytics dashboard which enhances the com-munication channels between content creators and users following strong ethical and legal guidelines, to answer the call for empowerment of independent creators in an inherently in-novative industry that is the one of online adult entertainment. This research thus provides a societal contribution just as significant as it sheds light on a field under-represented in the academic literature.

#

-Author: Harris Othman

#

-Supervisor: Asssitant Professor Dr. Leto Peel

#

-Second assessor: Assistant Professor Dr. Catalina Goanta

#

In case you have any questions with regard to this project, feel free to contact me via h.othman@student.maastrichtuniversity.nl or othman.harris@gmail.com .

#

-Steps (Topic Modelling (TM)) (Programming language: Python):

1. Transform the data for analysis (i.e., flatten the json file based on the number of comments per video)

  1.   #### Note: If your dataset already has an equal amount of rows for each comment per video, then you can skip this step

2. Clean the data (i.e., flag bots and non-English comments)

3. Fit a LDA model

3.bis Compute the Hellinger distance between topics to cluster them

4. Label the comments with a topic

#

-Steps (Sentiment Analysis) (programming language: R):

1. Clean the comments

2. Create a custom sentiment analysis dictionary

3. Compute the polarity score for each of the comments