An overview of the entire model-training process: data ingestion, model creation, model training, and model completion.
This series of illustrations was developed to help explain the Pienso process to new and prospective users. It chronicles the user journey, from initial data formatting and upload to model deployment and analysis (via an extended visual metaphor of Pienso-as-machine).
Data Formatting - Pienso accepts data files in csv, tsv, and json formats.
Data Ingest - A specialized upload portal allows you to refine your data set before training a model.
Model Creation - With your training data prepared for use, you’re just a few permutation choices away from creating your initial model.
Model Training - Training a model involves reviewing the topics that Pienso detects and deciding if they are worthy of enshrining with a label.
Model Output - Once you’re satisfied with the topics in your model, ‘finalize’ it to enable deployment and analysis.
Analysis - Completed models can be used to analyze new data sets or live data sources using the Data Set Analysis and Dashboard tools.
Classifiers - The individual topics within your finalized model can be deployed individually, for greater topic-sorting accuracy.
Deploy - Models can be deployed in external applications by generating a Pienso API key.
Pienso eventually moved on to a different illustration style to match our evolving brand identity, but these illustrations remain a useful visual aid for explaining Pienso to new team members.
