Problem

Background
The price of electricity in the Philippines is one of the highest when it comes to Asian countries [2] . Furthermore, its cost has been continuously changing and people are gravely affected due to price increases which disrupt their monthly budget flow, [3], especially when paired with the rising inflation and cost of daily necessities such as food in the Philippines. [4]. In fact, electricity costs have been continuously trending upward for the first three months of 2024 [5] which even goes counter to Meralco's earlier prediction of power rate reduction for March [6]. With the energy sector included as one of the current Marcos adminitration's main priorities [7], it is even more important to study the cost of electricity in The Philippines.

Questions
With this problem in mind, this study aims to answer the following main quetions:
1. How has household electricity price changed over the years?
2. Which factors (e.g. Transmission Charge, System Loss Charge) have affected the changes in electricity price the most?

Null Hypothesis
There have been no significant changes in electricity price over the past decade.

Alternative Hypothesis
There have been significant changes in electricity price over the past decade.

Action Plan
Use data science concepts to analyze past electricity costs (Meralco) to gain insights on the factors that most greatly affect it, which can help point out parts of the electricity production and distribution pipeline need to be improved the most.

Data Collection

Meralco Data Reports
We collected data found on the website of Meralco. Specifically, we collected most of their reports regarding the Typical Consumption Levels of their consumers via automated scripts , except for the May 2010 data which needed to be manually encoded and verified. The monthly reports gathered by the team gave data from January 2010 until March 2024 with only few missing data for the following months:
- September 2013
- October 2013
- November 2013
- January 2017
- August 2017
- October 2023
- November 2023
This allowed us to collect more than 2500 data items.
To have a more accurate comparison throughout the years, we also need data on the Purchasing Power of Peso (PPP) during the monthly time frame of our collected data. The PPP would allow for a more nuanced understanding of the impact of inflation on purchasing power by taking into account relative price levels between each time period in the dataset. The data was obtained from the OpenSTAT website of the Philippines Statistic Authority. The first dataset from OpenSTAT only contained information from January 1957 to November 2021 so a second dataset was needed to fulfill the time frame, January 2010 to March 2024, of our dataset.

Data Preprocessing
In order to combine all of the data in a single collection, we used a Python script to convert the data from the raw PDF tables to CSV files which are easier to process. Column headers and data values are kept exactly as is. The respective month and year for each data item added as new columns. The individual CSV data per month were then merged into a single dataset. Since some parts of the reports change over time, particularly columns are sometimes added, removed, or renamed, the merged dataset has empty values, which are kept as is. The changes in columns are also documented in the Python script.

The two Purchasing Power of Peso datasets came in CSV format. The first dataset used 2012=100 as the baseline reference year, while the second dataset used 2018=100. For a consistent PPP measure, the 2018=2010 dataset is converted to also use the 2012=100 baseline reference year using a Python script, which is possible since both datasets have data for overlapping months, specifically for July 2018, which allows us to compute the conversion factor. The two CSV datasets are then merged into one Purchasing Power of Peso dataset.

Data
You can view our data here.

Let's talk about our data science methodology.

Proin odio consequat sapien vestibulum consequat lorem dolore feugiat.

Here's what we found out.

Proin odio consequat sapien vestibulum consequat lorem dolore feugiat.