01
Overview
The PUV Modernization Program has been one of the relevant issues in the Philippines. Not only has it affected PUV drivers, but also commuters which encompasses the majority of the people. Numerous strikes were held in hopes of the government and drivers to have an agreement regarding this, which affected commuters. With this, we have decided to study the general reaction of the people regarding the issue which can provide insights on how the program is perceived by them.
02
Problem
Background
On June 19 2017, the Department of Transportation (DOTr) issued Department Order No. 2017-011, otherwise known as the Omnibus Guidelines on the Planning and Identification of Public Road Transportation Services and Franchise Issuance which launched the PUVMP (Public Utility Vehicle Modernization Program). [ 1 ]
The program aims to transform the road sector of public transport through the introduction of safer and climate-friendly vehicles, improved regulation, and industry consolidation. The program is beneficial in the long run however operators are caught in a difficult situation due to the high cost of the program. According to the Office of Transportation Cooperatives, jeepney operators must raise P300,000 in paid-up capital. Operators may have to spend between P20,000 to P30,000 each to comply with other requirements. Under the Omnibus Franchising Guidelines, cooperatives and corporations will also be required to eventually upgrade their fleet into modern jeepneys, which can cost up to P2.5 million per unit. [ 2 ] [ 3 ]
The program would surely improve public transportation however, jeepney drivers and operators – who already subsist on low wages – must then work to pay off millions in debt and interest.
Research Questions
- 1. What is the general reaction of the population regarding the PUV Modernization Program?
- 2. What are the frequent words used when the user activity increases?
Null Hypothesis
The ratio of positive to negative tweets is constant throughout the study period.
Action
Plan
Analyze social media sentiment through tweets to understand public opinion on the PUV Modernization Program.
Alternative Hypothesis
The ratio of positive to negative tweets significantly changes throughout the study period.
Keywords
Collection
Due to recent modifications to X/Twitter's API, the data collection tool SNScrape has exhibited limitations in retrieving data effectively. So instead, we employed a manual Twitter scraping approach targeting a minimum of 1000 tweets from 2022 to the present (2024) encompassing keywords directly relevant to the research topic.
Pre-processing
Collected tweets underwent rigorous pre-processing to ensure quality. This includes removing duplicates, non-text content, and formatting inconsistencies. The pre-processed data also underwent validation checks to ensure consistency and adherence to the defined criteria for inclusion.
Exploration
After collection and pre-processing, the data will undergo exploratory analysis to gain insights into tweet content, distribution of keywords, and potential presence of outliers. This initial exploration will inform subsequent data cleaning and analysis strategies.
04
Machine Learning
Extracting Topic Modeling Features using LDA (Latent Dirichlet Allocation)
In natural language processing, Latent Dirichlet Allocation (LDA) is a generative statistical model (unsupervised) that explains a set of observations through unobserved groups, and each group explains why some parts of the data are similar. The LDA is an example of a topic model. In this, observations (i.e., words) are collected into documents, and each word's presence is attributable to one of the document's topics. Each document will contain a small number of topics.
05
Results
Research Question 1
What is the general reaction of the population regarding the PUV Modernization Program?
Looking at the tweets from the range of months taken, the number of tweets with negative sentiments is evidently the highest out of the three classifications for each of these months. From this observation, we can safely say that the general reaction of the population regarding the PUV Modernization Program is mostly negative.
Analysis of the topic clustering shows that tweets concerning all of the topics are either mixed or negative. This implies that the negative sentiment is much more prominent compared to the positive and neutral sentiments.
Research Question 2
What are the frequent words used when the user activity increases?
Observing which months have the highest number of tweets, we found that March 2023, December 2023, and April 2024 have the highest user activity. From these three months the most frequent words from tweets with positive sentiments are modernization, jeepney, and PUV. The most frequent words from tweets with negative sentiments are jeepney, drivers, modernization, and notojeepneyphaseout.
Machine Learning Model
Implications
Learning the different perspectives and points between the tweet sentiments helps us understand the complexities of the PUV Modernization Program’s impact on various sectors of our country. The topics identified through clustering have shown both positive and negative significance, especially the ones with mixed sentiments. The significance of these topics include:
- The economic implications of the program have different short-term and long-term effects. The long-term may bring benefits, but the short-term could impose financial burdens.
- The modernized PUVs are said to be safer, more reliable, and environmentally friendly, an overall better alternative to traditional PUVs. However, the transition to this may lead to reduced availability of transportation options.
- The strong opposition to the phaseout may encourage policymakers to adjust the program accordingly, but may also cause social unrest or disruptions such as the effects of jeepney strikes.
Nevertheless, the tweets have weighed the pros and cons of the PUV Modernization Program, and most have decided that it would lead to a more negative outcome than what it promises itself to be.
06
Conclusion
The people have, no doubt, divided opinions towards the application of the PUV Modernization Program to the country. While being mostly negative, we must not disregard the number of positive and neutral sentiments as they may bring out essential points that contribute to the issue and to understand their perspectives.
Reviewing the collated tweets from using LDA, we determined these essential points that are in play:
- The program’s economic effects due to transition
- The phaseout’s public and transport implications
- The strong opposition to the phaseout
- The impact on the driver's livelihood
Limitations of the Study and Actionable Recommendations
A deeper analysis of the tweets is necessary due to the tweets being mostly non-standard Filipino, many contain double meaning or deeper connotation that the English translation may not capture. Using a natural language processing technique designed for the Filipino Language can help analyze the tweets better.
The lack of a standardized method for classifying the sentiment of tweets, intuition was used to determine the sentiment of the tweet which can cause bias. Developing and using a more standardized framework can help with sentiment classification.
Use of a more tuned ML model to accurately find the common topics among the tweets.
07
Who are we?
Justin Andrei
Gonzales
I'm Justin, a 2nd year BS Computer Science student at the University of the Philippines, Diliman. With my interest in Computers and Technology, I am currently finding my passion where I can utilize my skills and help the community.
In my free time, I enjoy watching K-dramas, playing video games like League of Legends and TFT. I also enjoy playing basketball whenever possible.
Godfrey Angelo Convento
Hey there! I'm Gelo, a 4th year BS Computer Science student at UP Diliman with a passion for front-end web development. I'm particularly interested in frameworks like Svelte and React.
Lately, I've been diving into the world of productivity systems, exploring tools like Todoist and Notion Calendar to keep myself on top of things.
When I need a break from the code, you might find me battling it out in DOTA 2 or Valorant. To unwind, I love jamming out to TWICE and BINI.
Angelo
De Leon
Hi! I'm Gelo, a 2nd year Computer Science student from the University of the Philippines. Having the interest for both programming and graphic design, I became passionate in honing my creative and technical skills to make a meaningful difference in the world.
Besides that, I mainly play games in my spare time. I enjoy games similar to Monster Hunter and Terraria, as well as some competitive games such as Valorant and TFT.