Time-Spotting

Abstract

Time-series forecasting is a crucial tool in various domains, such as finance, economics, and healthcare, to predict future trends and make informed decisions In this paper, we present time sight, an application that provides a comprehensive framework for time-series forecasting using the prophet library. Time-sight allows users to load and pre-process time-series data, configure the forecasting model with various settings such as horizon, seasonality, trend components, and holidays, fit the model and generate forecasts, evaluate and validate the predictions using cross- validation, perform hyperparameter tuning, and export the forecast, model retries, and configuration for further analysis. Our application provides a user- friendly interface and automates many of the pre-processing and modeling tasks, making it accessible to users with varying levels of expertise. We demonstrate the capabilities of time-sight using real-world datasets and discuss its potential applications and limitations. Our app can be used as a powerful tool for time-series forecasting in various domains, and we hope that it will enable more accurate and informed decisions based on future predictions.