Generating Notifications Through Chart Pattern Detection

A non-transitory computer readable storage medium embodying a computer program for performing a method, said method comprising:

• training a deep transfer model by:
• extracting historical charts from an analytics application;
• taking images of the historical charts;
• and
• associating domain-specific labels with images of the historical charts,
• storing a plurality of patterns in a repository, each pattern including metadata comprising a reference to a unique pattern name, a reference to a chart in a dashboard of an analytic application to be monitored, and a confidence threshold;
• scanning the dashboard of the analytics application;
• creating a chart image of a first chart;
• applying a deep transfer model to match the chart image with a first pattern of the plurality of patterns present in the repository based on the metadata;
• generating a classification vector from matching between the deep transfer model and the chart image;
• referencing a first confidence threshold of the first pattern to generate a notification from the classification vector;
• and
• communicating the notification to a recipient.

A computer system comprising:

• one or more processors;
• a software program, executable on said computer system, the software program configured to cause an in-memory database engine of an in-memory database to;
• store a plurality of patterns in a repository, each pattern including metadata comprising a reference to a unique pattern name, a reference to a chart in a dashboard of an analytic application to be monitored, and a confidence threshold;
• scan the dashboard of the analytics application;
• create a chart image of a first chart;
• apply a deep transfer model to match the chart image with a first pattern of the plurality of patterns present in the repository based on the metadata;
• generate a classification vector from matching between the deep transfer model and the chart image;
• reference a first confidence threshold of the first pattern to generate a notification from the classification vector;
• and
• communicate the notification to a recipient.