COLA 2023

This page contains the data and the plots used in the evaluation of the paper Dandelion: A scalable, cloud-based graphical language workbench for industrial low-code development (2023).

Process mining meta-model

• Meta-model definition

  • Event log meta-model

• Data generator

  • The following code inserts models of sizes until 1,000,000 into a designated Elasticsearch instance.
  • Edit the URL variable in src/createExperiments.ts. Then, run npm install and npm run createExperiments.
  • metamodelGenerator.rar

Measurements and graphs

RQ1: Loading large models

Can large synthetic models be loaded using Dandelion’s pagination scalability mechanism?

• Data: experiment1.xlsx
• Figure 17: Varying model size and page capacities. experiment1.r

RQ2: Pagination capacity

What is the practical limit of Dandelion’s page capacities?

• Data: experiment2.xlsx
• Figure 18: Varying page capacity on a fixed-size model. experiment2_bars.r
• Figure 19: The relation between page capacity and the number of nodes and edges. experiment2_scatter.r

RQ3: Injection of an existing low-code technology into Dandelion

Can we inject an existing industrial low-code technology into Dandelion, to visualise large industrial models?

• Figure 21: Meta-classes frequency in ROSE data. experiment3_bars.r
• Figure 23: Number of attributes and relations per meta-class in the resulting meta-model. experiment3_scatter.r
• Loading times: experiment4.r