Skip to main content

QuickStart: Analysis

Using aggregation pipelines and statistical functions to analyse your data.

Aggregation​

Aggregation summarises portions of data to give you an overview of what is stored.

Distinct Aggregation​

Get a unique list of values for any property:

//#region
// All unique IDs of public calendars
calendars = find unique _id from ${calendar:public}
print calendars
//#endregion
//#region
use training
// Unique dataset values in your private objects
datasets = find unique dataset from ${object}
print datasets
//#endregion

Aggregation Pipelines​

Pipelines let you filter, group, and sort while aggregating one or more fields:

//#region
use training
// Sum process executions grouped by status, sorted by quantity descending
summary = aggregate ${exec}
    match service = "ETL"
    group _id = "$status", qty = count()
    sort qty desc
end

print summary
//#endregion

Statistical Functions​

Simple Regression​

The simpleRegression function fits a linear model to a TimeSeries and can predict future values:

//#region
// Create a TimeSeries with 5 values
input = TimeSeries("2021-10-01", "DAILY", [12.5, 12.8, 12.9, 11.5, 11.9])

// Run regression
reg = simpleRegression(input)

print reg.slope
print reg.intercept
print reg.RSquare

// Predict the next day's value
print reg.predict(Date("2021-10-06"))
//#endregion
Statistical Function Library

OpenDataDSL includes an ever-growing library of statistical functions for TimeSeries analysis — including mean, stdev, min, max, sma, csum, and many more.

Next Step

In QuickStart: ETL you will extract data from remote sources, transform it, and load it into the platform.