Using the Cognos Analytics Assistant

Using the Cognos Analytics Assistant
Using the Cognos Analytics Assistant
Written by Mike Bernaiche, April 14th 2022

Have you ever wondered, “where do I start in building a dashboard in Cognos Analytics?”  Do you sometimes feel overwhelmed with all of the data?  I have and I bet you have as well.   What if I told you that IBM Cognos Analytics with Watson has a built in feature called the Cognos Analytics Assistant and it can help answer the questions above and so much more. 

What is the Cognos Analytics Assistant?

The Cognos Analytics Assistant is a text-based feature that allows you to gain quick insights into your data, making creating dashboards or explorations simple.  With just a few steps you can access data sources, view suggested visualizations or insights, and bring them over to a dashboard or exploration.  The assistant panel is made up of 3 sections: toolbar, output, and input areas. 

Cognos Analytics Assistant
Cognos Analyitcs Assistant Chat

Capabilities and Commands

There is a lot to unpack in this area about everything you can do with the Cognos Analytics Assistant.

  • Show Overview of Assistant – Type “help” to see a general overview of the assistant.
  • View Data Sources – To see a list of supported data sources type “view data sources.”
  • Suggest Questions – Type “suggest questions” and the assistant will suggest some starting questions for you.
Cognos Analytics Assistant Suggested Questions
  • Show Data Source Details – Type “show source (enter source name” or “load source (enter source name)” to see fields and details of your source.
Show Data Source Details
  • Show Column Details – type “show column (column name)” to see information and related fields.
Show Column Details
  • Show Column Influencers – Enter “What influences (column name)” to see what fields influence your chosen field.
Show Column Influencers
  • Create chart of visualization – The assistant can create a chart of visualization by simply typing “show chart (column1) and (column2)”.  You can further expand by then clicking “show related visualizations”.  You can further the experience by clicking “create dashboard form chart”.
Create chart of visualization
Related Visualizations
Create dashboard from chart
  • Create Dashboard – If you want to select a data source and ask IBM Cognos Analytics with Watson to start with a dashboard you can simply type “create dashboard” and use that as a starting point.
  • Create Simple Dashboard – With larger data sources or just the need for something simple, you can start with “create simple dashboard” to ask the Cognos Analytics Assistant to create a simple dashboard to start with.

Dashboard Example:

Cognos Analytics Assistant Dashboard Example

Filtering with Assistant

You can ask the Cognos Analytics Assistant to apply filters to data by applying words and associated synonyms.  I will go through a few examples below but there are many ways to filter through the assistant and a complete list can be found here. Assistant filtering - IBM Documentation

Filtering through the assistant has been included for many versions, however, time entity filtering is included with Cognos Analytics 11.2.2 which was released in March 2022.  To read about everything new in CA 11.2.2 please see this recent blog.  What’s New in Cognos Analytics 11.2.2 - Lodestar Solutions

General Filters

This filters examples are provided through IBM documentation.

  • show the top 3 products by sales
  • show sales where month is January
  • show products by sales where inventory > 500
  • show products by sales for California
  • show top <num> <column1> by <column2>

The following operators or modifiers can be used to filter.

  • In, for, if
  • Where, with
  • including
  • excluding
  • top percentage
  • bottom percentage
  • greater than
  • less than
  • top (post data aggregation)
  • bottom (post data aggregation)

Time Entity Filtering

The Cognos Analytics Assistant understands references to date, time, month, year, seasons, and holidays.  For the Assistant to apply time entity filters, the data set must have a Timestamp or Date/Time column.  The Assistant can filter based on relative time references (this, that, last, and next).  For example, show sales this Thursday.

The Assistant can use closed time intervals to filter.  A closed time interval has a start and end date. Example, show laptop sales for January 2021 or show sales from January 2021 to March 2021.

Other examples from IBM documentation are:

  • Show sales before last Thursday
  • Show sales in 2021
  • Show profit for December 31, 2021 at 10 pm
  • What is our expected revenue on the day before Labor Day?
  • Show me sales for camping equipment from June 1-5
  • Show me total sales last year
  • Show the average sales for past 3 weeks
  • Show the top 3 products sold on the third Monday after Christmas
  • show sales after next Christmas
  • Show October 2020 sales by product
  • Show sales after October 12, 2020
  • Show me sales up to a week ago
  • Show sales between last Monday and last Wednesday
  • Show sales every January and March
  • Show sales every January and every March
  • Show sales every single August
  • Show revenue in Canada for every Friday
  • Show sales each August and September and each Monday
  • Show sales by product each and every August
  • Show sales where transaction date is not January
  • Show sales where transaction date is not before January
  • Show sales where transaction date is not between January and March.

The following time-based filters are not supported:

  • References to time zones (ex. show sales for October 1 at 2 pm GMT)
  • Fractional periods (ex. half day or quarter month)
  • 2nd Tuesday
  • Weekly or monthly
  • Month over month, YTD, and QTD
  • Time periods such as Q1 and Q2 are not supported.

As you can see there are many ways to filter based on time, and If I were to take a guess the unsupported filters will be supported in due time by IBM.  I want to be clear I do not have any inside knowledge, but IBM has done an excellent job of continuing to update an already great product.

IBM has documentation on tips and best practices for using the Cognos Analytics Assistant that can be found here.  Check it out and take your knowledge to the next level.  Tips and best practices - IBM Documentation

Summary

As you can see the best place to start in Cognos Analytics may very well be the Cognos Analytics Assistant.  You gain great insights, start a dashboard, build a dashboard and so much more.  Simplifying using a few commands and natural language will get you started and wow your business.  If you need help upgrading to the latest version to take advantage of the time filtering or just general knowledge of the assistant please contact us at service@lodestarsolutions.com and we would be happy to talk with you. 

Forecast in Planning Analytics with AI

AI and Forecasting in Planning Analytics
AI and Forecasting in Planning Analytics
Written by Mike Bernaiche, October 14th 2021

Did you know that you can forecast in Planning Analytics Workspace by using automated tools that model time dependent data?  What does this mean for you?  You can take a series of time related data and let Planning Analytics suggest a forecast for the coming years based on statistics and trends from years past.  You can create a forecast, forecast low or forecast high version simply based on the trends of the past.  Below I will walk you through the steps to get started and create a forecast in Planning Analytics.

Requirements to use Forecasting

Time Dimension

  •  One or more time dimensions can be used to forecast against, and they must all appear on the column of the exploratory view.  In either case, the leaves must result in contiguous, evenly spaced periods of time.

Historical Data

  •  The historical data must represent at least double what is being forecasted.
  • The historical data must represent at least two seasons of data if it has seasonality.  For example, if the data cycles yearly and the leaf time members are at a month level of granularity, then it must have at least 24 months of data in the history.  This way, the feature can automatically detect the cycle.

Forecasting Data Cells

  •  The historical data must represent at least two seasons of data if it has seasonality.  For example, if the data cycles yearly and the leaf time members are at a month level of granularity, then it must have at least 24 months of data in the history.  This way, the feature can automatically detect the cycle.

Version Dimension

  •  The version dimension, if you want to use one, must appear in the context of the exploratory view.
AI and Forecasting in Planning Analytics

In the example above we have historical data to 2013 and will be forecast 2014 to 2019.

The Process to Forecast in Planning Analytics

The process starts by highlighting the members you want to forecast.  In the example below you will see that I selected all members shown.  Then click the forecast button .  This will open the forecast dialog box.  You will see 2 options.  Set up and advanced.  The first forecast we will create is in the set-up area.  I am forecasting years 2013 to 2019 based on the set up below. 

forecast button
forecast in planning analytics

Once you have selected everything correctly you will see the preview and forecast button enable and you can select one.  Preview will provide the below info.  You will see I changed the drop down to member 4 and the accuracy based on the AI in the Planning Analytics is high.

forecast preview

If you click the statical details tab you will see all the different statistical points.  This is for all of you data scientist out there!  You can click the learn more link and it will open a window that explains all of this information.

statistical details

If you like what you see you can simply close the preview window and click forecast.  The process will run and you will see all of the forecasted data in the cube.  You will notice the comment flag in 2013.  If you open the comments, you will see that the system has added in the statical information for each member.

forecasted data
comments

Advanced Tab Forecast

Keeping the time frame loaded in the set-up tab, you can now choose to have the forecast written to the optional version dimension (if you created one for your cube).  You can also ignore certain time frame(s).  This can be useful for times like the last 18 months of Covid data that may skew your data high or low, depending on your business.  Now you can select the version dimension and hierarchy.  And assign where you want the values to be loaded.  See below.

select version dimension and hierarchy

Now hit the forecast button and watch the forecast in Planning Analytics get loaded into the version dimension.  I moved the version dimension to the rows to show the below results.  As you can see Planning Analytics wrote data to the Forecast, Forecast Low and Forecast High fields.

forecast low and high

For detailed information on Forecast in Planning Analytics

Forecasting - IBM Documentation

For what is new in Planning Analytics Workspace

What’s New in IBM Planning Analytics Workspace Version 64, 65 and 66 - Lodestar Solutions

Summary

As you can see the process is very simple to get the automated forecast in Planning Analytics.  Take your planning process to the next level with this automated forecasting process.  Is it perfect, no.  But does it give you a great starting point with some artificial intelligence, yes!  For more information on forecasting or anything else planning analytics related reach out to us at services@lodestarsolutions.com

Are you ready to become a Transformational Leader?

Are you wondering how you become the transformational leader?  Lodestar Solutions is here to help!  You might want to check out our blog, It’s Time for CFOs to Leverage Transformational Planning.  

But more importantly Enroll in our complimentary webinar on Thursday, November 18th  from 10:30am-noon Eastern.

Mastering Transformational Planning: Why CFOs Should Lead this Initiative?

In this webinar you will discover:

  •  Why every CFO must discover and lead Transformational Planning
  • How to identify where the quick wins are in your organization
  • How to move your organization to Transformational Planning
  • How to find more time for Transformation Initiatives
  • And more….

To Register for this complimentary webinar, click here.

X