Macroeconomic Forecasting Using Python & Machine Learning (2010–2027)

A real-world data project that forecasts India’s GDP using key macroeconomic indicators such as Electricity Generation, Forex Reserves, Trade Balance, and Inflation. Built a machine learning pipeline using LinearRegression() with over 90% model accuracy, visualized economic patterns, and communicated insights for decision-makers.

🧠 Objective

Forecast India’s GDP for FY 2025–2027 using historical data (2010–2023) and analyze the impact of macroeconomic variables through data-driven modeling and visual storytelling.

📁 Data Overview

📂 Component	Details
📅 Time Period	FY 2010–11 to FY 2023–24
🔢 Variables	GDP, GVA, Electricity, Forex Reserves, Exports, Imports, Trade Balance, CPI, WPI, Industrial Index
🛠 Sources Used	Government reports, RBI datasets, Open Data API
📊 Dataset Format	Cleaned, merged, and normalized using `pandas`

✅ Cleaned and merged 10+ indicators

✅ Removed nulls, handled units, harmonized yearly values

✅ Built relationships between indicators using statistical metrics

🔍 Step 1: Exploratory Data Analysis (EDA)

✅ Identified strong correlations using Pearson’s coefficient
✅ Created year-wise trend plots to visualize indicator behavior
✅ Found outliers and consistent shifts using sns.pairplot(), heatmaps

Key Findings:

Indicator	Correlation with GDP	Observation
Electricity Generation	+0.92	Strong positive driver
Forex Reserves	+0.76	Moderately correlated, rising since 2014
Trade Balance (negative)	-0.42	Weak inverse relationship
CPI & WPI	+0.68 / +0.52	Inflation has mixed but noticeable impact

🧠 Step 2: Modeling and Forecasting

✅ Model Used: LinearRegression() from scikit-learn