💡EconInsight & 🔮EconViz

An Analysis of Global Digital Trade and Technology Adoption

Exploring the key drivers and leaders of the digital service economy from 2010 to 2023.

This repository contains the data, analysis notebooks, and Streamlit web application for the "Programming with Data" final project. The project investigates the landscape of global digital trade and examines the econometric relationship between technology adoption and digital service exports.

---

📋 Table of Contents

1. Key Research Questions
2. Data Overview
3. Technical Implementation
4. Project Structure
5. Usage

---

✨ Featured Visualization: Interactive 3D Data Cloud

To showcase the relationship between Internet Usage, GDP, and Digital Service Exports, we created a fully interactive 3D scatter plot using Plotly. The GIF below demonstrates the interactivity.

---

🚀 Key Research Questions

This project answers four primary questions:

1. 🌍 Global Digital Trade Landscape

Which countries lead in digital service exports and how has this evolved?

Takeaway: A small group of developed nations, including Ireland, the United States, and Germany, consistently dominate the top ranks. However, the animated bar chart race reveals a dynamic mid-tier, with countries like China and India demonstrating significant upward momentum over the last decade.

2. ⚙️ Technology Adoption & Trade

Is there a relationship between internet adoption and digital service exports?

Takeaway: Yes, but the story is complex. While a strong positive visual correlation exists, our rigorous econometric analysis shows this simple relationship is misleading. After controlling for country-specific characteristics and global time-based shocks using a Two-Way Fixed Effects model, the direct impact of internet usage becomes statistically insignificant, highlighting the importance of advanced modeling to avoid false conclusions.

3. 📡 The Digital Divide

Is the gap between developed and developing countries closing? How are adoption rates changing over time?

Takeaway: The gap is closing rapidly. The difference in average internet adoption between developed and developing nations has shrunk significantly, driven by high-percentage grwoth in the Global South.

Journalistic Narrative

Analytical Overview

4. 💡 Regional Growth Impact

Which regions are leading digital transformation? How are digital service exports growing by region?

Takeaway: The DiD analysis shows the established leaders are accelerating their lead. North America and Europe have a statistically significant, faster annual growth trend in digital exports compared to the baseline.

---

📊 Data Overview

The analysis is based on a comprehensive panel dataset compiled from primary public sources.

Variable	Description	Source
Exports_Digital_Service	Volume of digital service exports	World Bank / UNCTAD
internet_usage_pct	Percentage of population using the internet	World Bank
gdp_per_capita	Gross Domestic Product per capita	World Bank
population	Total population	World Bank
is_developing	Development status classification	World Bank

The final dataset covers approximately 150 countries over a span from 2010 to 2023. I utilized the wbdata Python library to systematically account for missing values in the main dataset.

Primary Data Source: World Bank Data (UNCTAD_DE)

The Raw Datasets

The Methodology of my project

---

🛠️ Technical Implementation

The project was executed through a series of robust data processing and analysis steps:

1. Data Cleaning & Harmonization

• Raw CSV files from the OECD and World Bank were ingested using Pandas.
• Datasets were cleaned, harmonized, and merged into a single, tidy panel DataFrame (final_panel_for_regression.csv).

2. Analysis & Modeling

• Baseline OLS: Initial regression analysis performed using Statsmodels.
• Panel Data Models: Implementation of Fixed Effects and Two-Way Fixed Effects models using the linearmodels library to control for unobserved heterogeneity.

3. Visualization & Dashboarding

• Interactive Web App: Built with Streamlit to allow users to explore the data dynamically.
• Advanced Charts: Choropleth maps, animated bar charts, and 3D scatter plots created with Plotly Express and Plotly Graph Objects.
• Custom Styling: A polished dark theme and custom CSS (via config.toml) ensure a high-quality user experience.

---

📦 Project Structure

KLDR-Project/

│
├── 📁 app/                   # Core Streamlit application source code
│   ├── 📁 .streamlit/        # Streamlit configuration files (e.g., config.toml for themes)
│   ├── 📁 data/               # Data files used directly by the live application
│   │   ├── 📁 output/         # Processed data from notebooks, ready for visualization
│   │   │   ├── 📁 question1/
│   │   │   ├── 📁 question2_paneldata/
│   │   │   ├── 📁 question3/
│   │   │   └── 📁 question4/
│   │   └── 📁 tables/         # Other miscellaneous data tables
│   ├── 📁 pages/              # Each .py file here is a separate page/tab in the Streamlit app
│   │   ├── 📄 data_assistant.py
│   │   ├── 📄 data_overview.py
│   │   ├── 📄 EconInsight.py
│   │   ├── 📄 EconViz.py
│   │   ├── 📄 homepage1.py
│   │   ├── 📄 homepage2.py
│   │   └── 📄 must-know.py
│   ├── 📁 utils/              # Utility scripts and helper functions
│   │   └── 📄 common_imports.py
│   └── 📄 app.py                # Main entry point to run the Streamlit application
│
├── 📁 drafts/                  # Folder for draft scripts or old code versions
├── 📁 scripts/                 # Standalone helper scripts for tasks
│
├── 📄 .gitignore               # Specifies files and folders for Git to ignore
├── 📄 README.md                # This project documentation file
├── 📄 requirements.txt         # List of Python libraries required for the project

---

💻 Usage

To run the dashboard locally:

1. Clone the repository:

   git clone https://github.com/kevindrmo/kldr-project.git
   cd kldr-project

2. Install dependencies:

   pip install -r requirements.txt

3. Run the Streamlit app:

   streamlit run app/app.py

---

_{Built by Kevin Di Raimo for the Data Analysis with Python Course of University of Zurich}