Pakistan-Economic-Dashboard

🇵🇰 Pakistan Economic Intelligence Dashboard

Interactive Power BI Dashboard + ARIMA Forecasting System

---

📌 Project Overview

A two-part Business Intelligence system built on real World Bank economic data for Pakistan (2000–2023):

Key Achievement: ARIMA model forecasts Pakistan's GDP, Inflation, and Population through 2028 — achieving 99.9% accuracy on Population and 88.4% accuracy on GDP.

---

🗂️ Repository Structure

Pakistan-Economic-Dashboard/
│
├── 📊 data/
│   ├── pakistan_economic_data.csv       # Main dataset (24 years × 16 indicators)
│   ├── pakistan_decade_averages.csv     # Decade-wise aggregated data
│   ├── pakistan_summary_stats.csv       # Descriptive statistics
│   ├── forecast_results.csv             # ARIMA actual + forecast values
│   ├── forecast_pivot.csv               # Wide-format forecast for Power BI
│   └── model_evaluation.csv             # MAE, RMSE, R², MAPE metrics
│
├── 🐍 scripts/
│   ├── STEP1_setup_and_run.py           # Master script: fetch data + run ARIMA
│   ├── fetch_worldbank_data.py          # Live World Bank API fetch
│   ├── generate_sample_data.py          # Offline data generator
│   └── task4_arima_forecast.py          # ARIMA(1,1,0) ML model
│
├── 📈 visuals/
│   └── forecast_charts.png              # ARIMA forecast visualization (3 indicators)
│
├── 📄 report/
│   └── Pakistan_Economic_Dashboard_Report.docx   # Full technical report
│
└── README.md

---

📊 Data Source

World Bank Open Data API — completely free, no API key required

https://api.worldbank.org/v2/country/PK/indicator/{INDICATOR}?format=json&date=2000:2023

Indicator	Code	Column
GDP (current US$)	NY.GDP.MKTP.CD	GDP_BillionUSD
Inflation (annual %)	FP.CPI.TOTL.ZG	Inflation_Pct
Population, total	SP.POP.TOTL	Population_Millions
Exports (% of GDP)	NE.EXP.GNFS.ZS	Exports_PctGDP
Imports (% of GDP)	NE.IMP.GNFS.ZS	Imports_PctGDP
Literacy rate (%)	SE.ADT.LITR.ZS	Literacy_Rate
Unemployment (%)	SL.UEM.TOTL.ZS	Unemployment_Pct

---

🚀 Quick Start

Prerequisites

Python 3.8+   # python --version
Power BI Desktop (free from Microsoft Store)

Step 1 — Run Setup Script

git clone https://github.com/YOUR_USERNAME/Pakistan-Economic-Dashboard.git
cd Pakistan-Economic-Dashboard
python scripts/STEP1_setup_and_run.py

This will:

• ✅ Install required Python packages automatically
• ✅ Fetch live data from World Bank API (or use embedded fallback)
• ✅ Run ARIMA(1,1,0) forecasting model
• ✅ Generate all CSV files + forecast charts

Step 2 — Open Power BI

1. Open Power BI Desktop
2. Home → Get Data → Text/CSV
3. Load: data/pakistan_economic_data.csv
4. Load: data/forecast_pivot.csv
5. Load: data/model_evaluation.csv
6. Build dashboard following /report/ guide

---

🤖 Machine Learning — ARIMA(1,1,0)

Why ARIMA?

Pakistan's economic indicators are non-stationary time series with consistent trends. ARIMA handles this via differencing, making it ideal for annual economic data.

Model Configuration

Parameter	Value	Meaning
p = 1	1 AR term	Current value depends on previous year
d = 1	1st-order differencing	Removes non-stationarity / trend
q = 0	No MA term	Keeps model simple for 24 data points

Training Setup

Train set:  2000–2019  (20 years = 83%)
Test set:   2020–2023  (4 years  = 17%)
Forecast:   2024–2028  (5-year projection)

Model Evaluation Results

Indicator	MAE	RMSE	R² Score	MAPE	Accuracy
GDP (Billion USD)	39.35	43.77	-0.116	11.58%	88.42%
Inflation (%)	7.25	10.32	-0.528	30.91%	69.09%
Population (Millions)	0.23	0.26	0.9966	0.10%	99.90%

5-Year Forecast (2024–2028)

Year	GDP (B USD)	Inflation (%)	Population (M)
2024	350.6	24.1	232
2025	362.1	22.8	236
2026	373.6	21.5	240
2027	385.1	20.2	244
2028	396.6	18.9	248

---

📈 Dashboard Features

Task-3: Interactive Dashboard

• 5 KPI Cards — GDP, Inflation, Population, Trade Deficit, Literacy
• Line Chart — GDP & Inflation trend 2000–2023 (dual axis)
• Bar Chart — Decade-wise economic comparison
• Pie Chart — GDP distribution across decades
• Data Table — Full raw structured data
• Area Chart — Exports vs Imports trade balance
• 2 Slicers — Decade dropdown + Year range filter
• Cross-filtering — Click any visual to filter all others

Task-4: Forecasting Dashboard

• Actual vs Predicted Line Chart — Real data vs ARIMA forecast
• Error Analysis Table — MAE / RMSE / R² / MAPE
• KPI Cards — Model accuracy + 2028 projections
• Forecast Image — Python-generated ARIMA visualization
• Type Slicer — Toggle between Actual / Forecast view

---

🔑 Key Findings

• 📈 Pakistan's GDP grew 5x from USD 73B (2000) to USD 375B (2022)
• 💰 Inflation peaked at 29.2% in 2023 — highest in 24 years
• 📉 Pakistan maintained a trade deficit every single year 2000–2023
• 📚 Literacy improved from 43.9% → 63.9% over two decades
• 🔮 ARIMA forecasts GDP reaching USD 396B by 2028 (+17% from 2023)

---

🛠️ Tech Stack

Tool	Purpose
Power BI Desktop	Interactive dashboard & visualization
Python 3.x	Data processing & ML modeling
pandas / numpy	Data manipulation
matplotlib	Forecast visualization
scikit-learn	Linear regression (AR coefficient estimation)
World Bank API	Live economic data source

---

📁 Files for Submission

✅ Task3_Pakistan_Dashboard.pbix          (Power BI — Task 3)
✅ Task4_Pakistan_ARIMAForecast.pbix      (Power BI — Task 4)
✅ Pakistan_Economic_Dashboard_Report.docx (Technical Report)
✅ scripts/STEP1_setup_and_run.py         (Python script)
✅ data/*.csv                              (All datasets)

---

👩‍💻 Author

Laiba Azha
MS Artificial Intelligence — PAK-AUSTRIA Fachhochschule
📧 laibaazhar.ds@gmail.com 🔗 https://www.linkedin.com/in/laiba-azhar-b89449263/

---

📜 License

This project is for academic purposes.
Data sourced from World Bank Open Data under Creative Commons Attribution 4.0.

---

Built with ❤️ using Python + Power BI | Spring 2026