In this project, we demonstrate the functionality of the SARIMA model (Seasonal Autoregressive Integrated Moving Average).
The dataset is sourced from Google Trends, where we analyze the popularity scores of specific search terms over time. The application allows users to upload new datasets for further exploration and analysis.
The first tab of the app serves as a data observatory. It enables users to explore trends by year and month, helping to identify seasonality and patterns in the data.
In the second tab, the SARIMA model is trained on the selected data. Users can experiment with seasonal steps and parameters to optimize the model's performance.
The core purpose of this project is to showcase SARIMA as a deterministic forecasting method and to highlight the potential of wrapping such techniques in accessible, interactive applications.
Download and run the Docker image of the project
You may visit the gihub repository of the projectt
Easy accesible data from Google Trends, try to download your desired data