Social Network Analysis on Indonesians’ Opinion upon Recession 2023 (Twitter Data)
What is Social Network Analysis?
Social network analysis (SNA) uses graph theory to understand and measure social networks.
It explains social behavior by analyzing connections, rather than looking at individuals in isolation. In simple words, we can understand the dynamics of topics happening in society or any other related information that might come from social connections, for instance from Big Data sources like Social Media.
In this article, we’d like to know the opinions of Indonesian digital society (from Twitter) in regards to the recession of 2023 that become a hot topic worldwide, in formulated question:
Do Indonesians have a positive forward-looking sight and have confidence in the upcoming 2023 challenges or have negative nuance towards recession?
Technically speaking, this work will be:
- Using Twitter Data by connecting through Twitter API Developer (free version) during 2 weeks of observations (6–23 Nov 2022)
- Keywords used in fetching tweets related to ‘Resesi’ (Indonesian language of Recession)
- All tweets use Bahasa as Indonesian Language (lang: id).
- Using Python 3.7 for SNA data preprocessing and for Network X (to draw the most discussed topics).
- Using SNA Tools: Gephi 0.9.7 to compute a Social Network Analysis graph.
Result: First Week Observation (6 Nov 2022–12 Nov 2022)
What are the most discussed keywords of recession?
If we filtered to be only keywords that have the most appearead among overall tweets from the first week of observation, here we see as following figure of Network X, the larger node (colored as orange round) means more appearance from overall tweets during the period:
Government, business, potential, UMKM (SMEs), optimism, growth, secured economy, a thread of inflation, PHK (lay off), etc.
Who’s got massive attention (Retweet, Quote, Reply) upon recession?
The top 3 accounts that obtained massive attention from the public through Retweet, Quote, and Reply are as following figure in the SNA graph (the larger font and rapid density of layers of a network mean the more immense attention they obtained from the public):
Which Tweets got massive attention (Retweet, Quote, Reply) upon the recession?
- @Okihita, with tweet:
2. @JukiHoki, with tweet:
3. @kozirama, with tweet:
Overall Sentiment
Here we see the first week of observation, society does not take economic recession as something serious to be taken care of, even further they believe the Indonesian consumption pattern will be resistant to the economic turbulence. Even more, we can see from the sentiment analysis gathered above, almost 80% of tweets are classified as neutral tweets, while positive 11% and negative tweets 9% among all tweets.
Result: Second Week Observation (15 Nov — 23 Nov 2022)
What are the most discussed keywords of recession?
Differing from the first week of observation, in the second week we see more serious talk about regional political nuance that shows keywords Ganjar, HIPMI, Governor, President, and current worldwide state such as economic impact, England, Russia, etc. Besides those keywords, trend #rimajuekonomitangguh also appear as one of the hot topic related to the recession.
Furthermore, it becomes interesting why the keyword ‘ganjar’ (Mr. Ganjar Pranowo is the Governor of Central Java who has served since 23 August 2013) appeared as one of the most talked topic, and if we deep dive into what’s happening on the grass root we can see a lot of tweets related to Mr. Ganjar such as:
Who’s got massive attention (Retweet, Quote, Reply) upon recession?
The top 4 accounts that obtained massive attention from the public through Retweet, Quote, and Reply are as following figure in the SNA graph (the larger font and rapid density of layers of a network mean the more extensive attention they obtained from the public):
Which Tweets got massive attention (Retweet, Quote, Reply) upon the recession?
- @catchmeupid with a tweet:
2. @BennyHarmanID with a tweet:
2. @restyca_yah with a tweet:
3. @ganjarpranowo with a tweet:
Overall Sentiment:
Here we see in the second week of observation, society has more in-depth talk more serious about what’s happening in the worldwide economy (such as in Russia and England’s recession) and also the nuance was shifted to regional politics. On top of that, from the sentiment analysis of 8095 tweets, 44% of tweets are classified as neutral tweets, while positive occupied 30% and negative tweets has 26% of all tweets.
Wrap-Up Result
The dynamics of topic shifting introduced from non-serious ones (more into archness or jokes) in the first week of observation, to more in-depth concern about economic worldwide and national political nuance in the second week of observation.
There was also sentiment shifting from the first-week observation to the second-week observation, that positive sentiment from tweets was increased from 11% to 30%, and negative tweets also increased from almost 10% to 26%.
All in all, if we gather two weeks’ observation the graph of Social Network Analysis is as follows, which tweets of non-serious ones obtained much more public attention rather than the challenges of the recession itself.
Ps: You can run the python code via Google Colab on your own, commencing by fetching the tweets through Twitter API Developer (if you have registered and obtained the credentials token) and preprocessing those tweets into Nodes and Edges to compute on Gephi Tools for Social Network Analysis graph.
Until next time, data enthusiast!
References:
- API Developer Twitter
- Alexander Kouznetsov and Maksim Tsvetovat (2011). Social Network Analysis for Startups. O’Reilly.
- Pacmann.io: Social Network Analysis
- https://cambridge-intelligence.com/
