Intro
Stacker is a highly configurable alternative data structure that transforms large amounts of data quickly and easily. It allows you to combine data from multiple platforms in such a way that multiple inputs can be defined as one single metric without the use of SQL or custom programming. Stacker also provides a detailed activity log to administrators, ensuring security and optimizing governance possibilities.
Using Enterprise Stacker, we can combine the following two tables that come from two different channels into one impressions table.
Using `Views` as the metric we are stacking, here is our output impressions table.
When should I use Stacker?
Stacker is a great option if you have any of the following situations:
Data exists in multiple platforms or DataSets Stacker makes it possible to combine and standardize metrics across platforms and visualization. |
Benefits include:
|
One Master DataSet is required or preferred Stacker combines data into one DataSet for use in visualization. |
Benefits include:
|
Data size is over 50-100M rows Stacker leverages Domo Query Performance. |
Benefits include:
|
Other benefits of Stacker include:
-
The output of Stacker is dynamic based on configured dimensions and metrics.
-
Unlimited input data sources.
-
Unlimited metrics/dimensions.
Tip: We recommend fewer than 50 metrics to optimize speed. -
Wildcard inputs - thousands of DataSets combined in minutes.
-
Optional View Output that can handle billions of rows of data, if needed.
Prerequisites
There are a few things you'll need to prepare and consider before using Stacker.
- Create a Service User Account with Admin access.
- Work with your IT or email administrator to generate a new email account for your organization.
- Assign that email a Domo license with Admin privileges.
-
Choose the naming conventions carefully. You'll need to make sure the names are all planned, consistent, and capitalized uniformly. Think through whatever is the most common or standard name for each particular metric in your business, and use whatever taxonomy you use internally.
-
Ensure the cleanliness of the data before stacking. Check the numeral and data values to ensure the columns include only numerical values.
-
Ensure you're applying the data type consistently (string, double, date, etc.).
How to get Stacker
- Contact your Account Executive or Technical Consultant to begin the process of deploying Stacker.
Note: Domo will access your instance to deploy the app and align it with the provided System User account's email. - Once deployed, Domo will provide you with the Stacker unique ID token. This token ensures an extra layer of security for your Stacker data. The first time you open the app, you'll input the token, and it will allow you to create your first Stacker job.
Using Stacker
Creating a new Stacker job
-
Select + New Configuration.
-
Choose which DataSet you would like to use for the metric.
Select a DataSet - Allows you to search by name to select one DataSet
Select all DataSets that match a rule - Allows you to select all DataSets where the name contains whatever string you specify.
Tip: When using the Select all DataSets that match a rule option, you must use all matching DataSets. For example, in the following image, all 43 DataSets will be used in the setup of that metric.
-
Name your metric and you can, optionally, set a metric alias. This is an alternative name you could use to search for this metric in your output DataSet.
-
Click Next.
-
Choose the value column that corresponds to the metric.
Note: If you must, you can choose a string. The Stacker engine will attempt to cast the string as a numeric value. However, if there are true string values in your column and it cannot be cast as a number, then you will receive an error after you run the job. -
(Optional) Create an aggregation on the value column. Select Next.
Note: When you choose to aggregate a value, Stacker will automatically group by all other non-aggregated columns that are included in the table. -
Choose a date column. This should be the date that lines up with the numeric metric you chose. You can select I don't have a date column if you don't want to specify one. Click Next.
Note: You can select a string type column here if needed. Stacker will attempt to cast the string as a date type. However, if there are values that aren't true dates in your column and cannot be cast as a date, then you will receive an error after you run the job. -
(Optional) Apply any necessary filters by selecting + add a filter. Select the column, operand, and value you would like to filter on. Filters provide the option for rolling windows of data, reducing data size, targeting specific numbers, targeting specific date ranges, etc. You can add as many filters as you would like. Click Next.
Note: Filters are applied with an "and" method. You cannot apply one filter or the other. -
(Optional) Add additional columns. You can create custom constant columns, select which columns are included, rename columns, and select data types
-
Custom constant column - Select
and type in the column name and choose a value you would like to be populated on every row for this DataSet.
-
Select DataSet columns - Choose which columns to include in the output DataSet.
-
Rename columns - You can rename any column you choose to include from the DataSet.
-
Select data type - Choose the data type you want the column to be in the output.
-
-
Select Next.
-
Name the output DataSet which will also be the name of this Stacker job.
-
Select an Update Schedule for the Stacker job.
-
Select Finish.
Adding multiple metrics to a Stacker job
-
Select the job you wish to add additional metrics to.
-
Click + Add Metric and follow the steps for Creating a new Stacker job.
Metric Actions
In the job configuration, you can copy , edit
, or delete
a metric.
Edit job configuration
To change the DataSet name or the job run schedule, select the Edit Configuration pencil.
Activity Logs
From the home screen, you can access the Activity Logs to see every action taken within the app and by whom.
Comments
0 comments
Please sign in to leave a comment.