Situation Tables Guide

Situation tables contain one row per situation (e.g., per order) and one column per feature of that situation. Such tables are commonly used as input for machine learning or statistical analyses of process data. In OCPQ, the bindings of a query box correspond to the situations, and its labels correspond to the features.

Defining Situations

First, follow the Query Guide to create a box whose bindings are the situations you are interested in. In our example, each binding consists of an order o1 together with its place order event e1 and its pay order event e2.

Adding Features

Next, add a label for each feature, as described in the CEL Labels Guide. For our example, we add two labels:

price, the price of the order (o1.attr('price')), and
time_to_pay, the time between placing and paying the order (string(e2.time() - e1.time())).

Exporting the Table

After evaluating the query, click on the number of bindings of the box. The binding table in the Inspector now contains a column for each label, and the expand button above the table shows it in a larger window. Using the export button 1, the table can be saved as a CSV or XLSX file.

Binding table with label columns

In the export dialog, you can choose

the format (CSV or XLSX),
whether to include the event and object IDs as well as a header row,
which labels to include, and
for boxes with constraints, whether to include the violation status of each binding.

Additionally, all attributes of the bound events and objects are added as extra columns.