Importing many batches / curves with meta information

Pro Importing batches / curves with meta info is only available in Visplore Professional.

In addition to numerical and categorical data columns, a unique feature of Visplore’s data model is representing curve-typed data columns in a table. Similar to "nested" arrays/lists in other environments, curve-typed columns hold an entire curve (=value vector) in each table cell. These curves can be visualized in Visplore Professional, and correlated with other information. This discrete data model has advantages for many use cases with large numbers of curves. Examples are:

This unique data model supports a joint analysis of the curves with meta-information, as described in detail in the "Analyzing many batches / curves" documentation chapter.
This chapter describes, how you can import a data table with curve-typed columns from various data sources. In particular, from Python, Matlab, and CSV files.

Importing curve-typed data from Python

To import curve-typed data from Python, the data needs to be in Pandas DataFrame format, where the cells hold the curves as nested lists. In the example below, Force and ForceSum are two curve-typed attributes:

Note: you can download a Jupyter Notebook featuring this dataset, to try this on your own.

To correctly import curve-typed data to Visplore, the format of the curve-typed columns need to meet the following requirements:

Once you have your data in the correct Pandas DataFrame format, you can send it to Visplore using the "send_data" command of our Python API. Note the following:

Some things you can do with curved-typed data in Visplore:

Importing curve-typed data from Matlab

Importing curve-typed data from Matlab has very similar requirements as the import from Python, thus please also see the section about Python above for the general workflow. In the following, only the Matlab specifics and the differences over Python are described.

In the below example, every cell of the table is a curve except the cells in the SimulationRunIndex and Plant columns.

Once you have your data in the correct format in Matlab, you can send it to Visplore using the "send_data" command of our Matlab API. Note the following:

Importing curve-typed data from multiple CSV files

You can also import curve-typed data to Visplore from CSV files, without Python or Matlab. For example, if you have one CSV file per operation pattern. To do this, you need a particular CSV file structure. First of all, you need a "master file" where you have one row per batch (or per operation, simulation run, experiment, ...). This "master file" can hold references to additional CSV files holding the curves of a particular batch/operation/simulation/experiment. Each additional CSV file must have at least 2 columns, one for the axis and one for the values of the curve.

Further details about the requirements of the CSV file structure:

Great! You have learned how to make use of Visplore's unique data model to import large numbers of curves! Please refer to our upcoming chapter of "Analyzing curve-typed data" to learn how to address typical use cases with this kind of data. As one example, labeling machine operations can already be found here.