Developed by Jozef Stefan Institute (JSI) as part of the enRichMyData toolbox, the Time Series Explorer Tool is focused on preprocessing and analyzing time series data. With capabilities such as data cleaning, de-noising, and advanced graph methods, it is built to handle challenges specific to time series datasets.
Key Features of the Time Series Explorer Tool
1. Descriptive Statistics:
Summarizes data with metrics like minimum, maximum, mean, median, quantiles, and autocorrelation functions using the pandas library.
2. Visualizations:
Offers plots like line plots, box plots, histograms, and seasonality plots
Includes partial autocorrelation plots for in-depth trend analysis.
3. Advanced Filtering:
Implements Kalman filters for predicting and de-noising data.
Supports Gaussian filtering to refine data for better predictions.
4. Change Detection:
Detects oscillations and replaces values with the mean of the region. Future updates may support median, minimum, and regression-based replacements.
5. Missing Value Imputation:
Imputes missing values using last-seen values and explores methods like interpolation and regression for better accuracy.
Service Openness
The Time Series Explorer Tool is available on GitHub at Time Series Explorer Tool