E-commerce websites get a continuous flow of useful data from various sources which can be analysed and utilised for the growth of the business. These data elements are referred to  as Streams that are collected at a destination source over the course of time. The stream is nothing but the set of events belonging to a particular data source typically in time series manner

In the case of the IQLECT’s Shopify App, the destination source is Ampere which is a platform created by IQLECT.

Each stream has a set of attributes that can be user-defined. The information that is collected in a stream can be utilised for creating charts, generate actions, link data sources, along with other functionalities wherever it is applicable.

The Streams Page that you see when you click on the Streams feature contains the list of data streams and derived streams in the system which are primarily related to the customers, orders, products, and page events on your online store. For example, the stream ‘Abandoned_carts’ contains attributes such as total_items_price, price_range, product_names, etc., that indicate complete information related to this stream.

There are basically two types of streams available on Ampere:

  1. Raw data streams: This is a basic stream with several attributes. Ex: CUSTOMERS.
  2. Derived streams: These are created on an existing stream using following functionalities:
    • Aggregation: Used to perform aggregation operations such as average, max, min, count and cardinality. Ex: Avg of Discounts-Price-Tax-Shipping.
    • Groupby: Performs operations such as average, max, min, count and cardinality with groupby on some attributes:
    • Static: These streams cannot be changed or modified later on.Ex: st_customers.
    • Static aggregation: Performs aggregation operations such as average, max, min, count and cardinality on a static stream. Ex: Abandoned_carts
    • Predictive: Used to apply and store the result of a given training model on a particular stream. Ex: (will be included when added on the app)
    • Complex Event Processing: Used to filter, combine, and correlate streams and to find state based patterns in continuous sliding windowed manner. The patterns identified with CEP are with 100 % confidence

To create a stream follow the steps given below:

1. Provide a Stream Name.

2. Click on Add Attributes. Enter Attribute Name, select Attribute Type, choose if the attribute should be indexed (check the box to enable indexing). Indexing will allow searching of data at a later stage.

3. Add more attributes if you want (by default a value of 1 is provided). By providing a numerical value in the field, several rows can be added.

4. Once you have added the required attributes, Click on Done.

5. Now, you have named the stream as well as the attribute. You can then click on Save to create the data stream which will be visible on the stream page on the right side.

The labels and fields associated with the streams are explained below:

Stream name: Name of the stream

Event count: Number of streams that have arrived in the last interval

Status: Shows whether the stream is enabled or disabled

Each stream has four quick launch buttons – Analyze, Visualize, Act, Drill which can be found in the row that contains the streams.

Launch Button Description
Takes the user to the aggregation wizard where they can modify the stream as per the requirements.

Takes the user to the chart creation wizard where they can modify the stream as per the requirements.

Act on stream
Takes the user to the actions wizard where they can modify the stream as per the requirements.

Drill down process
Initiates drill down process for the stream.

This feature is available only for the raw data streams. It creates a new stream with similar attributes and structure however the data input will be different in both the cases.

The checkbox ‘Show only data streams’ filters the entire list of streams and displays only the data streams by hiding the derived and/or analytics streams.