Hello. We are the Gold Diggers.
And this is our project, Analysis on Tweets Supposing the "Golden Era" during the FEM regime, where we dig through the reasons why the "Golden Era" is regarded as it is—hence, our name.
About Us
We are a group of BS Computer Science students from the University of the Philippines, Diliman. For our final project in our course, Introduction to Data Science, we prepared this study and presented it during Watch The Facts! A Mini-Conference on Data Science and Disinformation in the Philippines.
Carmelo Atienza
Hey! I'm Melo. I'm generally interested in emerging technologies and consumer tech. I also enjoy watching NBA (Warriors Bandwagon since 2016), playing RPGs and FPS games, and consuming anime and manga.
cratienza1@up.edu.ph | 09292942886
Pauline Abcede
Annyeong! I’m Pau, a BS Computer Science student sa umaga, Virtual Assistant of an American influencer sa gabi, and zombie sa madaling araw. In my free time, I’m your basic B–hanging out with friends, listening to KPop, watching KDramas and anime, and reading manga.
mpabcede1@up.edu.ph | 09760314172
Calvin James Maximo
Hi! I'm CJ. I like playing FPS and racing games (my faves are Apex and Forza). I play the guitar, and I am a bit fast at solving a Rubik's Cube (check out my WCA profile 2018MAXI03 :>) I also like watching K-Dramas and sports such as tennis, F1, and NBA.
ctmaximo1@up.edu.ph | 09272446167
Here's an overview of our Project.
The noise of misinformation and disinformation deafens the history and democracy of our country.
Motivation
With another Marcos seated in the position of President, the issue of historical distortion regarding the administration of former-President Ferdinand Marcos is more alarming than ever. As such, our team is motivated to contribute in any way we can in countering the dis/misinformation spreading rapidly through social media.
Problem
As we believe that education is the primary key in fighting dis/misinformation, it is important to identify which areas of learning we may need to work on more intensively to guide us in formulating a proper plan of action to counter the issue of historical distortion.
Solution
To find an answer to this question, our team plans to study dis/misinformative tweets implying how FEM's regime was a “Golden era” and identify which reason appears to be their most common claim. This should help us gain insights on how to approach our nation's problem on dis/misinformation regarding the FEM regime through education.
Research Question
What were the most common reasons in dis/misinformative tweets supposing the “Golden era” during FEM regime?
Hypothesis
The dis/misinformative tweets were most likely to claim a common particular reason leading to the supposed “Golden era.”
Null Hypothesis
The dis/misinformative tweets were equally likely to claim various reasons that led to the supposed “Golden era” during the FEM regime.
Solution
Collect dis/misinformative tweets, identify each tweet's main reason for their claim, and tally and rank their reasons according to frequency.
Data Exploration
Tweets were scraped manually with assistance from a third-party API scraper (Apify), prompted with keywords related to our topic. Then, we cleaned, tokenized, and applied Natural Language Processing to our data before using graphs to visualize our dataset. We then manually clustered each tweet according to their reason mentioned.
Machine Learning
To further explore our data, the group also attempted an approach in topic clustering using Machine Learning, on top of our manual clustering of tweets during the data exploration stage. The topics were clustered according to their implied reason for supposing the "Golden era" during FEM's regime, using the unsupervised learning model Latent Dirichlet Allocation (LDA) and t-Distributed Stochastic Neighbor Embedding (t-SNE).
.png)
Statistical Modeling
We will be using the Chi-square goodness of fit test to check if the distribution of our data are equal or not. This perfectly fits with our objective to find out whether there is a common particular reason standing out from the tweets or not.
Here are our key findings.
After exploring, testing, and analyzing our data, we have figured some important realizations.
01
Economy is ranked first among the reasons implied by the dis/misinformative tweets by manual labeling.
02
Statistical testing shows that there is a significant difference between the tweet distribution by reason.
03
Machine learning results show that the reasons being specified in the tweets are widely overlapping.
What's the bottomline?
Identify
Given our results, it implies that the supposedly good economy and quality of life during FEM's regime are the primary reasons why the "Golden era" is deemed as it is.
Improve
We first suggest extending the study by increasing the sample size and exploring other machine learning models and approaches to obtain more accurate results.
Inform
Once there is enough information, we can take more concrete actions on fighting dis/misinformation through education while taking our data into account in planning.
Our vision for this study is to be able to identify in which particular fields of knowledge do the dis/misinformed individuals about FEM's regime need enlightenment on. By educating them on these topics, we may be able to shed light upon the truth behind the spreading dis/misinformation. The findings may also help in revising lesson plans or maybe even improve the curriculum so that the younger generations may already absorb the right information at an early age, instead of having to correct them later on.
