Social media is one of the top distribution channels for user created content. By analyzing social media platform user shares, like tweets, photos, likes, re-tweets etc. one can see which types of posts receive the most engagement and use that rich information to form marketing, communication, channel and content strategy. We have a project and articles on social media analytics https://www.researchgate.net/project/Multi-Channel-Social-Media-Predictive-Analytics
Conference Paper BUSEM at SemEval-2017 Task 4A Sentiment Analysis with Word E...
Thesis Applying Machine Learning and Natural Language Processing Te...
Conference Paper Sector based Sentiment Analysis Framework for Social Media v...
However we would like to hear suggestions and alternative recommendations from the researchers who are working on social media analytics.
What are the algorithms / tools / software packages do you recommend for Social Media Analytics ? How do you relate your solution to machine learning or deep learning frameworks?
I propose an analysis of sentiment carried out on substantive content contained in comments written by users of social media portals.
This type of analysis is used, among others, as a method supporting research processes, the purpose of which is to determine predictions for specific phenomena and processes continued in the near future.
i recommend you to collect information using api or web scrapping technique and then for analyzing purpose you can use tools like matplotlib,seaborn etc
and algorithm??????????
i don't recommend any algorithm,most of the time simple bayesian theorems are used for such tasks but i will recommend you to spend more time on analyzing and cleaning your data,then you will know which algorithm can do well in your data for predictive analysis,i recommend you python as your main weapon,easy to learn,enriched in powerful libraries and most preferred for data scientists,i don't recommend you sentiment analysis,everyone is doing this,you can do clustering to seperate group of people,if your dataset is small then i will recommend you to try hierarchical clustering and if it is big then you can start with k-means clustering to seperate group of people,then you can do further EDA on those each group people,maybe you would like to give this competition a try too : https://www.kaggle.com/c/web-traffic-time-series-forecasting/data
The first and most important step is to collect the data: images, tweets, videos, etc. I know that YouTube have recently released video datasets to be used in Deep Learning applications. You can also look into Instagram to see if they provide any APIs to collect images. To collect tweets, I have used the TweePy Python library ( https://tweepy.readthedocs.io/en/latest/ ) easy and convenient. I have built an open-source, simple application to collect tweets, let me know if you need the link.
Dear Ayata,
You can use Netvizz for Facebook. For Twitter, I recommend the "rtweet" which is a package for the programming software R. Moreover, if you want to social network analysis, the Gephi is a free tool for it. By using "Twitter Streaming Importer" you can get the data as a network form in the Gephi. But if you want to get data from Twitter, you have to apply for a developer account to access Twitter APIs. Regarding the engagement; you can check on these studies which focus on the relationship between engagement and post category and post type:
Bonsón, E., Royo, S., & Ratkai, M. (2015). Citizens' engagement on local governments' Facebook sites. An empirical analysis: The impact of different media and content types in Western Europe. Government Information Quarterly, 32(1), 52-62.
Sobacı, M. Z., Hatipoğlu, İ., & Korkmaz, M. F. (2018). The effect of post type and post category on citizen interaction level on Facebook: The case of metropolitan and provincial municipalities in the Marmara Region of Turkey. Sub-national democracy and politics through social media (pp. 91-105). Cham: Springer.
Depends on what Social Media data you want to analyze. For Twitter I used http://twitter4j.org/en/, for Facebook https://restfb.com/ worked well for me. May be this paper is of interest:
Article An approach for big data technologies in social media mining
we looked at a number of tools that researchers use for social media analysis, and put together a little blog suggesting some people to follow, conferences where researchers talk about these, and shared the list of 100 tools we found: https://ocean.sagepub.com/blog/social-media-data-in-research-a-review-of-the-current-landscape
I also wrote a piece on how LinkedIn data is being used: https://blogs.lse.ac.uk/impactofsocialsciences/2019/07/09/using-linkedin-for-social-research/
And we had an intern that looked at the use of Weibo data: https://ocean.sagepub.com/blog/how-researchers-around-the-world-are-making-use-of-weibo-data
The data analysis, content of entries, comments on the portal social media are used in recent years combined technologies of advanced acquisition, storage, processing and data analysis involving the combination of technologies: - Big data database systems, - use of sentiment analysis on collected large collections of information obtained from the content of entries, comments contained on social media portal, - application of multi-criteria analytics of the type of computerized analytical platform Business Intelligence, - Besides, more advanced analytical solutions also include elements of machine learning and / or artificial intelligence technology.
You can follow
Article An approach for big data technologies in social media mining
Thanks for answers and comments Dariusz Prokopowicz , Syed Mobassir Hossen , Gharib Gharibi , İbrahim Hatipoğlu , Peter Wlodarczak , Daniela Duca , Osama Thamer Hassan , Moaid K Hussain
- Badges
- Science topic
- Similar topics
- Social Science
- Media
- Social Media