Strategy One
Time Series Trend Troubleshooting
-
Attribute and Metric Requirements For accurate predictions, make sure to use exactly one attribute and one metric in your questions to be placed in the X and Y axes, respectively. Queries not meeting this requirement might not yield optimal trends.
-
Break By Attributes: When you include a break by attribute to represent trends for multiple elements, such as analyzing sales revenue trends for the past 24 months for each store location, ensure the number of elements (store locations in this case) in the attribute are limited to enhance the readability of the analysis results.
-
Consider Attribute Forms The attribute used for trend analysis should have at least one form with a type of date, datetime, or integer. Timestamp data is ot supported.
-
Filtering DateTime attributes Filters on dates are not supported if the attribute a DateTime type.
-
Smart Attributes Trend analysis with smart attributes is not currently supported.
-
Match data type between database and Metadata Trend analysis on data that is stored as an Integer in a data warehouse, but defined as a Date attribute in metadata is not supported.
-
Consolidation and Grouping Trend analysis against consolidated or grouped attribute elements is not possible at this time.
-
Cartesian join Trend analysis is not performed when a cartesian join is detected in the dataset(s).
-
Additional information about trend Follow-up questions about the underlying trend algorithm in Auto Answers are not supported. This includes questions such as, "What is the intercept value?" To get more information about the statistical properties of the trend analysis, use the tooltip for visualizations the info icon in Auto Answers.
-
Derived objects Trend analysis cannot be performed by Auto Answers if the requested metric is not present in an underlying dataset and needs to be created as a derived metric on the fly.