writeup.Rmd

---
title: "writeup"
author: "Rizka Nugrahaeni"
date: "`r Sys.Date()`"
output:
  word_document: default
  pdf_document: default
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```
## ### Write-up for Final Project: Exploring Financing and Climate Outcomes in Asia

#### **Research Question**
The primary aim of this research is to explore the relationship between financing commitments and their impact on renewable energy adoption and CO2 emissions in Asia. Key questions include:
- How do financing commitments influence renewable energy consumption?
- Is there a measurable relationship between financing and a country’s share of global CO2 emissions?

#### **Data and Methods**
##### **Data Sources**
1. **ADB Climate Change Financing Data (2020):** Provides details on financing commitments by country and sector.
2. **World Bank Data (2021):** Includes renewable energy consumption and CO2 emissions from the power industry.
3. **Worldometers (2021):** Offers global CO2 emission shares for each country, retrieved via web scraping.

#### Data Wrangling: Libraries and Functions Used

1. **Data Wrangling Libraries:**
   - `tidyverse`: For data manipulation and visualization.
   - `rvest`: For web scraping the CO2 emission data.
   - `stringr`: For text cleaning and formatting.
   - `httr`: For managing HTTP requests.

2. **Key Functions:**
   - `read_csv()`: Import datasets from CSV files.
   - `mutate()` and `select()`: Transform and extract relevant columns.
   - `pivot_wider()` and `pivot_longer()`: Reshape datasets for analysis.
   - `inner_join()` and `left_join()`: Merge datasets.
   - `replace_na()`: Handle missing values.
   - `html_node()` and `html_table()`: Extract and clean web-scraped tables.
   - `str_to_title()` and `trimws()` to align country names across datasets..

##### **Analysis Approach**
1. **Data Visualization:** Created grouped bar plots and scatter plots to examine the relationship between financing, renewable energy adoption, CO2 emissions, and global emission shares.
2. **Regression Analysis:** Fitted linear models to evaluate the relationships:
   - Financing commitments vs. renewable energy adoption.
   - Financing commitments vs. CO2 emissions.
3. **Interactive Visualizations:** Developed two Shiny apps:
   - A choropleth map for visualizing emission shares and financing commitments.
   - A dynamic bar plot comparing financing, renewable energy, and emissions by country.

#### **Results**
##### **Static Figures**

1. **Bar Plots:**
   - Countries with higher financing commitments (e.g., China and India) have significant renewable energy and CO2 emissions contributions, but the relationship isis   inconsistent.
   - The share of global emissions for smaller countries is often negligible, overshadowing their renewable energy progress.
   - Grouped bar plots compare financing commitments, renewable energy consumption, and CO2 emissions.
2. **Scatter Plot (Financing vs. CO2 Emissions):**
   - Scatter plots highlight relationships between financing and global CO2 emission shares.
   - A weak correlation is observed, with outliers like China contributing disproportionately to global emissions despite large financing.
3. **Choropleth Map:**
   - A geospatial map visualizes the share of global CO2 emissions, highlighting regional disparities.
   - Visualizes regional disparities, emphasizing China's dominance in both financing and emissions contributions.
4. **Sentiment Analysis**
    The project includes a sentiment analysis of an article from AsianInvestor.net about climate finance. Using the `tidytext` library, key positive and negative terms were      identified, revealing public perceptions of climate finance.
    - Positive sentiments in the article highlight themes like "support," "sustainable," and "innovative."
    - Negative sentiments focus on risks and delays, indicating skepticism about the efficiency of climate finance mechanisms.

##### **Regression Analysis**

1. **Renewable Energy Model:**
   \[
   \text{Renewable Energy} = \beta_0 + \beta_1 \times \text{Total Financing} + \epsilon
   \]

   Where:
   - \(\beta_0\): Intercept (base level of renewable energy with zero financing)
   - \(\beta_1\): Effect of total financing on renewable energy consumption
   - \(\epsilon\): Error term

2. **CO2 Emissions Model:**
   \[
   \text{CO2 Emissions} = \beta_0 + \beta_1 \times \text{Total Financing} + \epsilon
   \]

   Similarly, the coefficients \(\beta_0\) and \(\beta_1\) describe the relationship between financing and CO2 emissions.

### **Output**
Two linear models were developed:
1. **Renewable Energy Consumption:**
   - Dependent Variable: Renewable energy consumption percentage.
   - Independent Variable: Total financing commitments.
   - Result: Weak positive relationship with limited explanatory power (Adjusted R²: 0.0059).
2. **CO2 Emissions:**
   - Dependent Variable: CO2 emissions.
   - Independent Variable: Total financing commitments.
   - Result: Minimal relationship, suggesting the need for additional explanatory variables.


##### **Shiny Applications**
##### **Shiny App 1: Bar Plot Comparison**

- **How to Use:**
  - **Variable Selection:** Users can select one or more variables (e.g., financing commitments, renewable energy consumption, share of global CO2 emissions) using checkboxes.
  - **Country Filter:** A dropdown menu allows filtering by specific countries of interest.
  - The results are displayed as a grouped bar plot, enabling direct comparisons across countries and variables.

##### **Shiny App 2: Choropleth Map**
- **How to Use:**
  - **Variable Selection:** Users can toggle between viewing "Share of World Emissions (%)" and "Total Financing Commitments."
  - **Opacity Slider:** Adjust the map's visual opacity to customize viewing preferences.
  - Hover over a country to display detailed values for the selected variable.

#### **Challenges**
- **Data Limitations:**
   - The dataset is limited to single-year observations, constraining trend analysis.
   - Missing data for some countries required imputation or exclusion.
- **Explanatory Power:**
   - The models show weak relationships, highlighting the need for additional explanatory variables (e.g., GDP, energy mix).
- **Web Scraping Challenges:**
   - Differences in country names across datasets required manual correction.

#### **Conclusion and Next Steps**
- Financing commitments demonstrate a limited direct impact on renewable energy adoption and emissions reductions.
- The weak correlations underscore the importance of policy frameworks, technological advancements, and country-specific contexts in driving climate outcomes.
- Future research should:
   - Expand the dataset to cover multiple years for a longitudinal analysis.
   - Include additional variables like GDP, energy policy indices, and energy mix.
   - Explore advanced modeling techniques to better capture country-specific effects.