DEVELOPMENT OF YOUTUBE SCRAPER AND YOUTUBE STATISTICS ALONG WITH MODEL UTILIZATION IN BANDUNG INSTITUTE OF TECHNOLOGY'S SOCIAL NETWORK ANALYSIS APPLICATION
Bandung Institute of Technology's Social Network Analysis Application (SNA ITB) is an application developed by Artificial Intelligence Center Bandung Institute of Technology. This application is capable of analyzing scraped data from several social media. Among the social media available on thi...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/85474 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Bandung Institute of Technology's Social Network Analysis Application (SNA ITB) is an application developed by Artificial Intelligence Center Bandung Institute of Technology. This application is capable of analyzing scraped data from several social media. Among the social media available on this application scraper are Twitter and Facebook. The analysis that can be carry out by the application is the classification of emotions, sentiments, hoaxes, depression, and hate speech.
The addition of the YouTube scraper feature is carried out using the Application Programming Interface or API provided by YouTube. The API is called Youtube Data API v3. The scraped data will be inferred by the models available in the application, namely sentiment, emotion, hate speech and depression. Data which has been successfully scraped and inferred is then stored in a database and then used by SNA ITB's statistical API to be displayed on the crawler statistics page and project statistics page.
Calculation of the accuracy of inference is carried out on data scraped from YouTube on emotion and sentiment models. Calculations were carried out by taking a random sample of 52 out of 200 comments from 6 videos. Calculations are carried out by comparing manual labels with randomly drawn samples.
The development of the YouTube scraper feature and its statistics was successfully carried out with unit test coverage of 87%. The full score coverage test failed to be achieved due to a failure in the Twitter API used and errors occurred in the Facebook scraper used. The results of calculating the accuracy of the inference results showed that emotional accuracy was 57.6% and sentiment was 66.2%.
The result of this final assignment is that the SNA application has new features, namely a YouTube scraper and YouTube statistics. Users of the SNA ITB application can scrap YouTube and analyze data from YouTube's comments related to emotions, sentiment, hate speech and depression. |
|---|