Detecting anomalies in visualizations

A processor readable non-transitory storage media that includes instructions for managing visualizations of data, wherein execution of the instructions by one or more processors, performs actions, comprising:

• providing a plurality of visualizations based on data from one or more data sources, wherein each visualization includes one or more marks that are associated with one or more values from the one or more data sources;
• determining one or more monitored visualizations from the plurality of visualizations based on one or more metrics associated with the one or more monitored visualizations;
• sampling one or more mark values from each monitored visualization, wherein the one or more sampled mark values are stored with a timestamp that corresponds to a time of that the one or more mark values are sampled;
• and
• in response to an amount of the one or more sampled mark values for a monitored visualization being greater than a threshold value, performing further actions, including:
• training one or more mark models to classify a portion of the one or more sampled mark values associated with the monitored visualization, wherein the one or more mark models predict one or more ranges of values for the one or more classified mark values associated with the monitored visualization;
• and
• in response to the portion of the one or more classified mark values associated with the monitored visualization having a value that is outside of the one or more predicted range values, providing an indication that the monitored visualization is associated with an anomalous mark value, wherein the indication is reported to one or more users.

A system for managing visualizations, comprising:

• a network computer, comprising:
• a memory that stores at least instructions;
• and
• one or more processors that execute instructions that perform actions, including:
• providing a plurality of visualizations based on data from one or more data sources, wherein each visualization includes one or more marks that are associated with one or more values from the one or more data sources;
• determining one or more monitored visualizations from the plurality of visualizations based on one or more metrics associated with the one or more monitored visualizations;
• sampling one or more mark values from each monitored visualization, wherein the one or more sampled mark values are stored with a timestamp that corresponds to a time of that the one or more mark values are sampled;
• and
• in response to an amount of the one or more sampled mark values for a monitored visualization being greater than a threshold value, performing further actions, including:
• training one or more mark models to classify a portion of the one or more sampled mark values associated with the monitored visualization, wherein the one or more mark models predict one or more ranges of values for the one or more classified mark values associated with the monitored visualization;
• and
• in response to the portion of the one or more classified mark values associated with the monitored visualization having a value that is outside of the one or more predicted range values, providing an indication that the monitored visualization is associated with an anomalous mark value, wherein the indication is reported to one or more users;
• and
• a client computer, comprising:
• a memory that stores at least instructions;
• and
• one or more processors that execute instructions that perform actions, including:
• displaying one or more of the monitored visualization or the indication on a hardware display.