Genereate realistic synthetic login events in order to train an ML model for fraud detection. Similar to Incubation, I've added it to the https://adosiml.com system to train fraud detection models alongside my customer recommmendations models, churn, segmentation, etc.
Example image output below; The CSV is only 18.1 MB with 56,000 rows. I tried a more enterprise value for 'login events each day' and the CSV got too large to even upload. When this is deployed as a class library within a deployable system, it will be much easier to train with ~1,000,000 rows for example.
CSV for ML data ingestion:
