Time series analytics evaluate the values of a given set of variables over time and group those values into a window for analysis and aggregation. The Vertica Analytics Platform is a scalable, fast solution for time series analytics. The optimized structure and the analytical capabilities of the platform, as well as the columnar nature of Vertica allows time series data to be sorted, compressed, and partitioned to enable optimal performance.
What’s more, Vertica provides some critical capabilities that make time-series easier to manage and analyze. For example, Vertica provides gap-filling functionality, which fills in missing data points, as an interpolation scheme. This is a method of constructing new data points within the range of a discrete set of known data points. The platform interpolates the non-time series columns in the data (such as analytic function results computed over time slices) and adds the missing data points to the output.
Event-based windows functions are part of Vertica’s standard SQL analytics. These functions simplify the detection of events in time series data. Event-based windows let you break time series data into windows that flag on significant events within the data. This is especially relevant in financial data where analysis often focuses on specific events as triggers to other activity. For example, given an input stream of stock quotes, the stock analyst may want to place the input quotes into a new group whenever the spread (the difference between the ask price and the bid price) exceeds $0.05. If we view each such group as a window of events, then the window endpoints are defined by the occurrence of certain event types.
|Analysis||Supported in Vertica|
|Event series JOINs||✔|
Machine Learning (Prediction)
Machine Learning (Categorization)
Event series joins
Vertica supports typical data warehousing query joins. The platform also provides the INTERPOLATE predicate, which allows for a special type of join. The event series join is an Vertica SQL extension that lets you analyze two event series when their measurement intervals don’t align precisely—such as when timestamps don’t match. These joins provide a natural and efficient way to query misaligned event data directly, rather than having to normalize the series to the same measurement interval.
Vertica natively supports path and pattern analysis through an event series pattern matching extension. The SQL MATCH extension lets you screen large amounts of historical data in search of event patterns. You specify a pattern as a regular expression and can then search for the pattern within a sequence of input events. MATCH provides subclauses for analytic data partitioning and ordering, and the pattern matching occurs on a contiguous set of rows.
In addition to time-series analysis, organizations are applying predictive analytics to everything from improving machine uptime to reducing customer churn. With Vertica, analysts can now leverage SQL to natively create and deploy machine learning models based on larger data sets without down sampling to accelerate the decision making process.