This project was developed as part of a statistical and machine learning analysis focused on forecasting UK vital statistics using time series modelling techniques.

The objective of the project is to analyse historical demographic trends and forecast future values for one of the following UK population indicators:

• Births
• Deaths
• Marriages
• Divorces

The selected dataset for this project focuses on:

The project applies statistical time series forecasting methods to identify historical patterns, seasonality, long-term trends, and future projections using real-world government demographic data.

A UK government statistics agency aims to improve demographic forecasting and policy planning using predictive analytics and time series modelling.

The task involves:

• Exploring historical demographic trends
• Cleaning and transforming time series data
• Identifying trend and seasonal behaviour
• Building forecasting models
• Evaluating model performance
• Forecasting future demographic outcomes

• R Programming
• Time Series Analysis
• Statistical Modelling
• Forecasting Techniques
• Data Visualisation

Libraries Used:

• forecast
• ggplot2
• tseries
• dplyr

• Imported historical demographic datasets
• Cleaned and transformed time-based variables
• Handled missing values and inconsistencies

• Visualised historical birth trends
• Analysed seasonality and trend patterns
• Identified fluctuations and long-term movements

Implemented forecasting models such as:

• ARIMA
• Exponential Smoothing
• Trend and Seasonal Decomposition

Evaluated model performance using:

• RMSE
• MAE
• Residual Analysis

Generated future forecasts for birth statistics in England & Wales.

• Identified long-term demographic trends in birth statistics
• Observed seasonal patterns and fluctuations over time
• Built predictive models capable of forecasting future population-related indicators
• Demonstrated the importance of statistical forecasting in public sector planning and decision-making

• Time Series Forecasting
• Statistical Analysis
• Data Cleaning & Transformation
• Exploratory Data Analysis
• Predictive Modelling
• Data Visualisation
• Analytical Storytelling

├── data/
├── plots/
├── Task3__Time-Series-Modelling-Births-England-Wales.R
├── README.md

• Compare additional forecasting models (Prophet, LSTM)
• Incorporate external socio-economic indicators
• Deploy forecasting dashboard using Streamlit or Shiny
• Automate model retraining pipelines

Deepa Ramu

MSc Data Science | Data Scientist | Machine Learning & Analytics Enthusiast

LinkedIn: linkedin.com/in/deepa-ramu03 GitHub: github.com/DeepaSaru