Social Media Data
Social media data mining collects and analyzes unstructured information (such as posts, comments, tweets, shares, likes, mentions, and clicks) shared on networks like Facebook and Twitter. This kind of information provides a rich source of text data for academic research in a variety of creative ways:
Identify current and future trends
Capture major social issues
Understand public opinions and sentiments on social issues
Analyze marketing and journalistic practice
Facilitate sophisticated predictive modeling
The Springer's article "Social Media Analytics: A Survey of Techniques, Tools, and Platforms" provides an easy-to-read overview and essential foundation knowledge of social media data mining.
According to another article "Collection Social Media Data for Research", there are three major ways to collect social media data:
Do it manually -- It is very time-consuming, and thus not recommended
Use web services -- It is the easiest way to collect data, but costs money. Also, you may not be able to get the data you want.
Use coding -- It provides the highest level of flexibility, but you have to learn how to code and keep updated on the changing policies of social media platforms
This guide will briefly discuss the second and third ways.
Use Web Services
Increasing number of web services are available on the market to help you collect and analyze social media data. Some services are developed with researchers as one of their primarily customers, including Volunteer Science and uMaxData. Some are developed for more general purposes but also commonly used by academia, such as ExportData and WebTunix.
Most web services can provide custom-made mining solutions to meet the specific requirements of sophisticated research projects, but they cost money. As an alternative option, the Library has subscribed to a web service plan (with specified data coverage) for the use of the HKBU community.
UMiner INSTALLED IN DAR OF MLC
The tool has been installed in Data Analysis Room. Please book a workstation in advance. Every user can create a personalized account so that previous search queries, search results and analysis can be saved and downloaded.
Sources -- Online news, Facebook and Twitter pages of KOLs and public sectors, public blogs and forums, etc.
Region -- Hong Kong data only
Coverage -- Jan 1 2017 - current
Data Mining Dashboard -- provides display of trends, source distribution, communication paths, automatic sentiment analysis, word clouds, top posts, etc.
Computer Assisted Content Analysis Module -- support data pooling, data categorization, sentiment analysis, etc.
Use Coding
If you wish to start with coding, you will need to know APIs, R, Python, etc. to do data collection. Please refer to the following resources to learn relevant techniques:
R FREE
Learn R FREE
Basics of R programming | No prerequisite knowledge
Python FREE
Data Software Training Videos: Python (Cantonese) DEVELOPED BY LIB
Data Software Training Videos: Python (Putonghua) DEVELOPED BY LIB
This video series provides basic programming concepts on Python codingIntroduction to Using Twitter Social Media Data in R (Lessons 1-3) FREE
Collecting and analyzing Twitter data with R | No prerequisite knowledgeWeek 2: Collecting and Extracting Social Media Data FREE
Collecting Twitter and YouTube data with Python | No prerequisite knowledgeHow to Use Facebook Graph API and Extract Data Using Python FREE
Collecting Facebook data with Python | Prior knowledge of basic Python is needed
Miscellaneous Tools
There are other social media research tools that can be used to collect (and/or analyze) social media data. Some tools are free to use and require little or no programming. Some tools are costly or coding intensive. The following lists provide an overview.