Where is a Python Software package used for analyzing geodetic data. Currently, the software only supports VLBI data, but work is underway also for other techniques. The Where software is made available to the geodetic community as an open source, MIT-licensed project.
Where is developed at the Norwegian Mapping Authority and made available under an MIT license. The MIT license is permissive allowing you to use, modify and distribute the software, as long as copyright and license notices are preserved.
Acknowledgement and Citing
If you are using Where for work or research, in presentations or as part of a code or analysis project, we ask that you acknowledge or cite Where as appropriate.
For information on how to get started using Where, see the Quickstart tutorial.
Where is built as a modular software so that it is easy to customize how the analysis done, change which models are used, and even write new models. The following describes some key concepts of the Where software.
A pipeline is a description of a particular analysis. Currently, Where only supports the VLBI pipeline which can analyze VLBI data. Through configuration profiles the VLBI pipeline can be set up to analyze different kinds of VLBI sessions, including R1/R4- and intensive sessions.
An analysis is a pipeline applied to specific data. Typically, an analysis is the result of running Where on one VLBI session. The product of an analysis is a dataset containing all the input and output of the analysis. In addition, further products like sinex files, estimation parameters, maps etc may be written at the end of an analysis.
Each Where pipeline has a configuration which defines the details of how the analysis is done. Examples of configuration settings for VLBI include
- which observation format to use (NGS or vgosDB)
- how to clean the data before analysis
- which site displacement models to apply
- which reference frame to use for a priori station coordinates
- which parameters to estimate
For each analysis, the pipeline configuration is copied and can be tailored to the specific session that is being analyzed, for instance by specifying clock breaks, or that observations for a given radio source should be discarded.
In order to simplify configuring Where for different kinds of VLBI sessions,
several configuration profiles are defined. The default configuration for
VLBI is tuned for R1/R4 24-hour sessions. The same configuration does not work
well for intensive sessions, for example, due to there being much fewer
observations so that only a few variables should be estimated. This is supported
by using the
intensives configuration profile.
Both input data and results from the modeling are stored in a Where dataset. A dataset essentially consists of corresponding time series for all relevant variables in an analysis.
A dataset can be explored interactively in the companion There software. In addition, Where can transform a dataset to other useful products such as SINEX files, estimated parameters or maps quantifying the result of an analysis.