Shadkoech/streamlit-financial-analysis

Machine learning–based financial analysis and forecasting dashboard built with Streamlit to analyze macroeconomic indicators, predict trade volumes, and simulate economic shocks across Sub-Saharan African economies. # streamlit-financial-analysis ## Overview `streamlit-financial-analysis` is an interactive financial analytics and forecasting dashboard built using **Python and Streamlit**. The project applies **machine learning, simulation techniques, and explainable AI** to analyze macroeconomic indicators and predict international trade volumes in Sub-Saharan Africa. The goal of this project is to provide an intuitive platform for exploring how macroeconomic variables such as **GDP, inflation, exchange rates, and foreign direct investment** influence trade dynamics across selected countries. The system integrates historical macroeconomic data and uses predictive modeling to forecast trade volumes while allowing users to simulate economic shock scenarios. Through an interactive Streamlit dashboard, users can explore economic trends, compare machine learning model predictions, and visualize insights that can support economic analysis and policy decision-making. --- # Project Objectives The main objectives of this project include: - Analyze historical **trade volume trends** across selected Sub-Saharan African countries - Identify **macroeconomic drivers** of trade performance - Build **machine learning models** to forecast future trade volumes - Cluster countries based on their **economic characteristics** - Simulate potential **economic shocks** using Monte Carlo simulations - Provide **explainable insights** using SHAP and LIME --- # Features The dashboard provides several analytical capabilities: ## 1. Exploratory Data Analysis - Visualization of macroeconomic indicators - Trade volume trends over time - Correlation analysis between economic variables ## 2. Machine Learning Forecasting The project implements multiple models to predict trade volumes: - Random Forest - XGBoost - LSTM (Long Short-Term Memory Networks) ## 3. Economic Clustering Countries are grouped based on macroeconomic patterns using: - K-Means Clustering - Density-based clustering methods …