Best Practice Guide

The ultimate step-by-step guide for analytic content creators.

Selecting the right type of automation

  • Threshold alerts

    Threshold alerts is an automation feature for dashboard reports. Users can configure rules in individual reports to define precisely when it should be sent based on predefined parameters. For example, your users can set a report to be sent based on specific metrics falling below a certain number. Alerts are then subsequently triggered based on the users’ configured parameters to automatically notify them when an event has occurred and to take action. Not all BI platforms include threshold alerts or automation as part of their toolset, so do your research if you consider these features as essential for your users’ day-to-day dashboard workflow.

  • User initiated (assisted insights explain + compare)

    User-initiated insights, also called assisted insights, is a BI feature powered by AI and natural language generation (NLG) that helps explain to dashboards users what, how and why certain metrics and results have occurred through automatically generated explanations and comparisons. In modern BI platforms like Yellowfin, this is an embedded dashboards feature.

    <insert graphic>

    Directly within your dashboard, chart and graph data can be explained or compared with another data point by user initiation. By right-clicking the relevant visualization, you can auto-generate charts and visualized narratives while on the same dashboard. All results are ranked by relevance to the user’s query and provide deeper detail into the what and why of your question. Before running auto-analysis on your dashboards, validate the embedded reports and metrics are up-to-date and cleaned to ensure relevance and accuracy in the generated results.

  • Advanced insight detection with Signals

    Yellowfin Signals offers automated business monitoring for your dashboards and reports to continuously monitor, search, and analyze your data to detect any significant changes and inform users of what, when, where and why something has happened so you can act fast. This advanced insight detection feature-set is exclusive to Yellowfin BI dashboards.

    Signals runs various machine learning algorithms and performs time series analysis to find the most interesting, important and relevant anomalies, disruptions and trends in your dataset. It scans all data defined in a given view, whereas standard alert-based reporting is typically only set up for an individual report with simplified rules, such as comparing a single value to a defined threshold. Yellowfin Signals goes a step further to allow you to deploy a number of sophisticated algorithms to identify a range of relevant changes in the dashboard dataset. Examples include total and average, trend direction, volatility, step shifts and outliers. 

    Yellowfin Signals automatically detects and accounts for natural seasonal variation in data, and allows for both a streamlined setup and fine-grained configuration of thresholds and algorithm parameters for deeper analysis. Both simple and advanced setup offers personalized results, based on your data permissions and how you interact with Signals. All alerts come with the natural language explanation and additional analysis of Assisted Insights on correlated data changes to help your user uncover the root cause of the alert immediately.

    In order to get the most out of its insight detection capabilities, let the system know what data trends and insights are important to you by creating a Signals analysis schedule, which can be updated as your data changes and scheduled to run the analysis as frequently as you prefer.