JDemetra+ is an open source software for seasonal adjustment and time series analysis, developed by the National Bank of Belgium with the support of the Bundesbank, Insee and the partners of the COSA, (COfidentiality Seasonal Adjustment and time series), project co-founded by Eurostat. JDemetra+ has been officially recommended by Eurostat to the members of the European Statistical System since 2015. It is unique in its combination of very fast Java routines, a Graphical User Interface and an ecosystem of R packages (rjdverse). The graphical interface provides structured visual feedback suitable for refined analysis and training. R tools allow to combine the capabilities of JDemetra+ with the versatility of the R world, be it for statistical functions or data wrangling. A pdf version of this documentation is available here

Main Features

JDemetra+ provides algorithms for:

Seasonal Adjustment
Trend estimation
Benchmarking and temporal disaggregation
Nowcasting
Revision analysis

Currently available versions

The latest release in version 3 family can be found here.

Version 3.x family provides, among other things, extended features for seasonal adjustment and trend estimation, including high frequency data.

The latest release in version 2 family can be found here.

The highest version using Java 8 is v2.2.4.

Installing the software

The sections below detail how to install JDemetra+ (Graphical user interface and Cruncher) and how to configure R to run rjd3 packages.

Graphical User Interface (GUI)

You can find the latest releases

for the v3.x family: here

You should install the latest release available denoted from now on v3.x.y

Scroll down the page, download and unzip the file jdemetra-standalone-3.x.y-windows-x86_64.zip, if you use Windows, or the zip corresponding to your OS in the list.
To start the application, run the file nbdemetra64.exe located in the following subfolder …\jdemetra-standalone-3.x.y-windows-x86_64\nbdemetra\bin

Version 3.x requires Java 21 or higher, jdemetra-standalone-3.x.y-windows-x86_64.zip contains a portable version of Java, so you don’t have to deal with this issue on your computer.

R packages related to version 3.x (rjd3…) also require Java 21 or higher, you can (and should) use the portable version provided with the graphical user interface to run them, this is explained here.

for the v2.x family: here

To install v2.2.4:

Scroll down the page, download and unzip the file jdemetra-2.2.4-bin.zip.
To start the application, run the file nbdemetra64.exe located in the following subfolder …\jdemetra-2.2.4-bin\nbdemetra\bin

Remark: You can create shortcuts to the executable files if you want to launch them from another folder (Desktop, project folder…).

Additional Plug-ins

To benefit from extended features of the graphical user interface installing additional plug-ins is required. The list of available extensions and the installation procedure are detailed here. How to access the added features in the GUI is described here.

In the v2.x family some of the additional features are:

Benchmarking and Temporal disaggregation Plug-in here
Nowcasting Plug-in here

In the v3.x family some of the additional features are:

Benchmarking and Temporal disaggregation Plug-in here
Seasonal adjustment of high frequency data Plug-ins here
Additional algorithms for seasonal adjustment Plug-ins here

Cruncher

JDemetra+ has an executable module, called the Cruncher, allowing to automate tasks

To install it :

in v2, go to the JWSACruncher page, download and unzip the compressed folder jwsacruncher-2.2.x-bin.zip, corresponding to the version of your Graphical User Interface.
in v3, go to the jdplus-main page, download and unzip the compressed folder jwsacruncher-standalone-3.x.y-windows-x86_64.zip, corresponding to the version of your Graphical User Interface.

R packages

JDemetra+ algorithms can be accessed in R, which is detailed here

You can directly head over to this GitHub page, then for each package indications for installation and basic use are provided in the readme files.

Configuration needed to run rjd3 packages

To use rjd3 packages in R, you need Java 21 or higher.

You can (and should) use the version of Java that comes with the Graphical User Interface (GUI). It is contained in the file jdemetra-standalone-3.x.y-windows-x86_64.zip available on this page, in the “Assets” section, which you might have to expand to be able to see all the files.

Once unzipped, the Java will be in located in C:\Software\jdemetra-standalone-3.x.y-windows-x86_64\nbdemetra\jdk-21.0.2+13-jre (path to be adapted to your computer/server): this location must be declared in R.

After unzipping the file:

run this code at the beginning of your programs

Sys.setenv(JAVA_HOME = ".../jdemetra-standalone-3.x.y-windows-x86_64/nbdemetra/jdk-21.0.2+13-jre")

declare it “permanently” is the Renviron file (where environment variables specific to R are stored), following the steps below:
- in the R console run file.edit("~/.Renviron")
- in the file that opens add the line :
```
JAVA_HOME = "C:/Software/jdemetra-standalone-3.x.y-windows-x86_64/nbdemetra/jdk-21.0.2+13-jre"
```
- save the file and restart R

Follow R packages installation procedure and run basic examples from the readme files of each package. They are all listed on this page

User Support and Useful links

If you need user support, please raise an issue in this repository

For specific questions related to production of seasonally adjusted series you can look up these pages

To keep up with all JDemetra+ related news head over to the JDemetra+ Universe Blog

You can also browse and watch tutorials for JDemetra+. On this YouTube channel you will also find JDemetra+ related webinars.

Background

This website is under construction, in the meantime you can fill a large number of the gaps by referring to the previous version of the on-line documentation, coordinated by Sylwia Grudkowska-Kubik (National Bank of Poland).

Eurostat’s recommendations on the statistical processes described in this documentation are outlined in:

Key methodological explanations and state-of-the-art description and references can be found in:

COSA project

JDemetra+ is developed and maintained in the framework of the Eurostat co-founded COSA (COfidentiality Seasonal Adjustment and time series) project.

It is composed of two Centres of Excellence (CoE): one on Time Series Analysis (TSA), coordinated by Insee (France), and one on Statistical Disclosure Control (SDC) coordinated by CBS (Netherlands). In a context of development of common tools and methods within the ESS, this action will carry on the work of the STACE project (2020-2024) around three pillars: knowledge sharing and support to practitioners, development and maintenance of bespoke software (JDemetra+ for TSA, τ-ARGUS and sdcMicro for SDC) and methodological guidance (Handbooks, Guidelines, research papers). The overarching goals is to provide user friendly, efficient and technically up-to date tools for statistical production in both fields.

More information on the project’s website https://cosa-grant.org/

How to contribute

If you want to help us improve this book, you can fork this repository on GitHub and create pull requests with your contributions.