Revision Date: 12 September 2022
This documentation has been created for software version 2022.3.0.0.
Suite 105, 220 Davidson Ave, Somerset, NJ, USA 08873
For more information, visit
Copyright © 2022 whiz.ai All Rights Reserved.
Trademarks owned by whiz.ai
No Warranties and Limitation of Liability
Every effort has been made to ensure that this document is an accurate representation of the features and functionalities of the WHIZ.AI platform. However, the development of the software is a continuous process. So, small inconsistencies may occur.
We would appreciate any feedback on this document
Send comments via email
WhizAI supports the following browsers:
- Google Chrome, v103 & above
- Mozilla Firefox, v103 & above
- Microsoft Edge, v103 & above (Windows); v102 & above (MacOS)
- Safari, v14 & above (MacOS)
This document outlines the new product updates for the release version 2022.3.0.0. The following lists the new features and enhancements.
Pinboards, Cards, and Responses
- Updating action fields on cards in pinboards
- Viewing data or images using external links on cards
- Drag and drop filters on pinboards
- Viewing timeline visualization for multi-dimensions
- Switching base metrics on cards
- Enhanced tabular responses for Period over Period (POP) and Comparison response
- Scrolling on pinboards
- Viewing smart totals
- Multiple selected values in user defaults
- Getting Insights into WhizAI
- Using Moving Annual Total (MAT) expression in queries
- Using enumerated dimensions in queries
- NLQ support: Best/top n brands for each region/territory
- Data not available on cards
- Using Half to Date (HTD) time expression in your queries
- Adding Half to Date (HTD) as a time period filter on the PoP card
- Adding multiple dimensions on a card having Period over Period (PoP) data
- Introducing the Script Editor
- Airflow 2.2.5 support for the Data Modeler
- Accessing the logs for the pipeline run tasks
- Configuring Data dictionary: Move numeric data from Dimensions tab to Metrics tab
- Pipeline configuration - Login credentials
- Adding a new pipeline: Selecting the Data process mode
- Introducing Narratives
- Managing custom narrative templates
Pinboards, Cards, and Responses
Updating action fields on cards in pinboards
Now, on pinboards, certain cards can be configured so that they show an actionable metric column.
You can update the actions performed against a particular record from the dropdown lists in this column.
For example, if you ask, ‘Show me top customers by status’:
The response displays a 'Status’ column.
When you hover the cursor over any field in this column, a tooltip is displayed asking you to select an action (No, Not Applicable, Yes) from the drop-down list, as shown in the following figure:
After you click Save, the updated actions get captured on the card.
This helps collaboration between the board users as all the board users can view the updated action for a particular record/s. Also, the board users can edit the action, if required.
- When the actionable column associated with the dimension is displayed on multiple cards, and you update action on a single card, the other cards will not be updated automatically even after refreshing the browser.
- If you want to view the actionable column in the response for the metric, then you have to mention the metric in NLQ. For example, ‘Show me customer by TRx status’, if you ask ‘Show me top customer’ then the actionable column is not displayed.
- When you apply a new metric filter on a card, then the available editable metric gets replaced with the newly added metric.
- For a card having a single metric, data available on the actionable column is not displayed for any other visuals (except for the table visuals).
Viewing data or images using external links on cards
Now, on pinboards and responses, certain cards contain clickable links that are associated with corresponding dimension values displayed in the cards.
For example, if you ask a query ‘Region by TRx by call goal’ then WhizAI shows the following response:
For any dimension value, when you hover the cursor over this icon WhizAI shows the following options:
- View link: redirects to an external link that shows the data.
- View image: redirects to an external link that shows the images.
When you click the link, you get redirected to a new tab where you can view data or images associated with those dimension values. This provides you with additional capabilities to perform a better analysis of the data displayed in the response.
- For the nested dimension response, when the hyperlinks are associated with both the dimensions, then except for table visuals the hyperlinks are accessible only for the nested dimension.
For example, when you ask ‘Show me, Customers, by Top 2 products’ then in the response, the hyperlinks are accessible in the table visuals for both the dimension Customer and Product. When you switch to other visuals, the hyperlink is accessible only for the Product and not for the Customers.
Drag and drop filters on pinboards
Now, you can drag and drop the filters on pinboards to reorder them.
For example, let’s say you add Region, Geography, and Products as filters on a pinboard
You can simply drag and drop these filters to reorder them, as shown in the following figure:
Viewing timeline visualization for multi-dimensions
WhizAI shows a timeline visualization on both pinboards and responses for the multidimensional queries.
For example, when you ask ‘Show me region by response dates’ then in the response you can see the option for timeline visuals as shown in the following figure:
- In the timeline response, if you export more than 20 data points, then the downloaded image gets compressed, which affects the quality of the image.
- Follow-up action for single or multi-dimensional timeline response is disabled.
- On pinboards and responses, the response date is not supported for a period filter.
Switching base metrics on cards
Now, on pinboards, cards with computation having the same base metric, if you switch the base metric by applying filters then it gets changed and computations remain the same way from the previous metric.
For example, when you ask WhizAI ‘Show me TRx growth and Trx market share by customers’ then you get the response for the metric TRx for the computations Growth and MarketShare as shown in the following figure.
Now, if you change the metric TRx to NRx, then the base metric for Growth TRx and TRx MarketShare will change to Growth NRx and NRx MarketShare.
Only the metric TRx will get changed to NRx and the computations Growth and MarketShare remain the same way as shown in the following figure. This helps in maintaining a single card for the base metrics which have common computations.
Enhanced tables for Period over Period (POP) and Comparison responses
Now, we have enhanced the tables shown in the POP or Comparison responses; you can now hide or show the available data in the tables.
For example, for a POP response if you ask a query, ‘Show me Brands POP for TRx, NRx’ then in the response, you can see a Columns dropdown list. From this dropdown, you can select or clear the checkbox to hide or show the data available on the response.
Similarly, for a comparison response, when you ask a query, ‘Compare Boston MA vs Chicago IL vs Philadelphia by months’ then in the response, you can see a Columns dropdown list. From this dropdown, you can select or clear the checkbox to hide or show the data available on the response.
The following enhanced table features are also available for the PoP and Comparison responses:
- Scroll bar to navigate to the last row
- Ability to search data in a dimensional column
- Pin columns to the Left/Right/No pin
- Option to simply copy and paste the data or copy with Headers Details
- Auto sizing Columns
- On cards, the sorting is not supported on the metrics columns for POP and Comparison response. For example, if you ask a query ‘brands pop for trx’ then you cannot not see the sorting options for metrics column.
- For POP and Comparison response, when you change base entity the data on the card also gets changed. For example, if you ask a query ‘boston vs chicago by customer’ and change the base entity the data on the card also gets changed.
Scrolling on pinboards
Now, when there are filter/s added to pinboards, the filters toolbar gets hidden while scrolling up or down the board; however, the filters toolbar is displayed back on your screen as soon as you stop scrolling.
Before scrolling on the pinboards the filters toolbar is displayed as shown in the following figure.
While scrolling on the pinboards, the filters toolbar is hidden as shown in the following figure.
Viewing smart totals
Now, WhizAI gives an ability to view totals for computed metrics for all the applicable responses. The totals are calculated based on computational calculations.
For example, if you ask WhizAI ‘Show me region by brand for TRx marketshare’ then in the response table, you will see the total percentage of each brand against all the regions.
Multiple selected values in user defaults
Now, WhizAI has the ability to show multiple default values for filters on pinboards. Thus, when you add dimension filters and select multi-select option, the user default values are selected for that dimension filters.
- The dimension filters will not be shown on the explorer for defaults with multiple options.
Getting Insights into WhizAI
Now, WhizAI provides an ability to derive meaningful insights from the data you want to analyze.
WhizAI offers the Insights Workbench UI to:
- Find top driving factors for a given business metric
- Detect Anomalies in a trend
There are two ways to find the top driving factors and detect anomalies.
- You can go to the ‘Insights Workbench’ UI offered by WhizAI and then perform a set of tasks that will help analyze the data.
- You can navigate to the 'Insight Workbench' from an applicable response by clicking the 'Explore Insights' option. In this approach, the details are automatically picked up from the response and pre-populated in the ‘Insights Workbench’ UI.
The ‘Insight Workbench’ interface has the following two tabs:
- Anomalies: Helps identify the anomalous behavior of your business metric.
- Key Drivers: Helps find out the top driving factor of your business metric.
- Anomaly detection and Key Driver analysis are not supported on a computed metric.
- Adding the same dimension as filter as well as factor is not supported for the key driver analysis.
- For the key driver analysis, the potential list of factors selected for the analysis must be part of the dimensionality of the metric to be analyzed. In case, one or more metrics are selected as the potential factors then ensure that the rest of the dimension factors are applicable for all the metrics including the metric to be analyzed. In case this condition is not met then the system would not generate the result and throw appropriate error message.
Using Moving Annual Total (MAT) expression in queries
We have enabled support for NLQs that contain Moving Annual Total (MAT) expressions.
MAT is the total sales value for last twelve months, as the twelve-month period moves forward with each month, the sales value from the latest month is added and the sales value from the oldest month is removed from calculation.
If MAT expression is included in the query, the response shows sales value for the last twelve months with respect to the time period mentioned in NLQ. If you don't specify the time period, then by default, the last twelve months are considered.
For more information, refer to the following examples.
Show me MAT sales trend for 2021
Sales trend for current twelve months for the year 2022 (Jan 2021 to December 2021).
What are my sales for MAT Jan 2021?
Sales for current twelve months including the month of January 2021 (Feb 2020 to Jan 2021).
Example: What are my sales for MAT Jan 2021?
Using enumerated dimensions in queries
We have enabled support for NLQs that render enumerated responses. An enumerated response contains a list of all the data assets or shows the total number of data assets for a dimension included in the NLQ. Data assets can be employees, resources, or patients etc.
Now, you can add and configure a data model, where you can configure the source data with reference to such dimensions and WhizAI provides enumerated responses.
Example NLQ: How many employees are COVID affected?
NLQ support: Best/top n brands for each region/territory
We have enabled support for following type of NLQs:
- Show me best brand for each territory/region
- Show me top 2 brands for each city
- Show me top 3 brands for each state
- Show me top 2 regions for each brand
For example: If you ask, ‘show me top 2 brands for each state’, WhizAI shows the response in a grouped way where top 2 brands from each state are available under each state. Refer to the following response.
Example NLQ: Show me top 2 brands for each state
Data not available on cards
Now, after applying filters on cards, if there is no data available for the selected filter values, WhizAI shows a corresponding message in gray color.
For example, if you select TRx as the metric and ProRata as the computation, and data is not available for this filter combination, then WhizAI shows the following message:
‘No data available for current context’
Using Half to Date (HTD) time expression in your queries
WhizAI supports the ‘Half To Date’ (HTD) time expression in NLQs. When you include HTD expression in a query, WhizAI provides a response with the metric value for the period starting from the beginning of the current semester till the current date. For more information, refer to the following example.
Example NLQ: HTD PoP
Adding Half to Date (HTD) as a time period filter on the PoP card
Now, on cards in pinboards, you can apply the HTD time period filter as shown in the following image.
Adding multiple additional dimensions on a card
Now, on the cards, you can add multiple additional dimensions from the Columns dropdown. In this case, the card displays rows for every combination of primary dimension with each of the additional dimensions.
For more information, refer to the following figures.
- Additional multiple dimensions are not supported on the cards having nested table responses.
Introducing the Script Editor
These custom scripts can be used for:
For example: For building a calculated metric. Market share can be calculated from existing metrics to get the market share of a product with respect to the other products.
- Application plugins
For example: Write a script to get condition-based custom responses for an NLQ.
From the Script Editor UI, you can:
- Add and manage custom scripts, that is, you can add, rename, edit, and delete the scripts
- View all the existing scripts (custom scripts and system scripts).
Airflow 2.2.5 support for Data Modeler
We have enabled Airflow 2.2.5 support for the Data Modeler.
Accessing logs of pipeline run tasks
Now, as an admin user, you can run a data pipeline and view the logs of this pipeline run task on Airflow.
Configuring Data dictionary: Move numeric data from the Dimensions tab to the Metrics tab
While configuring the data dictionary, after configuring all the dimensions, from the 'Actions' column, you can move numeric data (rows) with the 'numeric' data type to the 'Metrics' tab. (Earlier ‘Move to Metric’ option was showing for ‘non-numeric’ data (rows) as well).
Pipeline configuration - Login credentials
Now, admin users no longer need to provide the user ID and password during the pipeline configuration, before running a pipeline. User ID and password are now configured in the common configuration as a part of product installation.
Following new configuration parameters are introduced:
Adding a new pipeline: Selecting the Data process mode
For database connections, we have added support for the ‘Batch’ process mode. Now, you can select Batch or Live as process modes while adding a new pipeline.
For Batch data process mode, all connections will be available to select from.
For Live data process mode, only database connections will be available to select from.
Batch mode: When a pipeline is configured and run, the system loads the data from the source system to the destination (Druid). All connection types now support data process in batch mode.
Live mode: The system has connectivity to the database, the system creates a live connection with the data source, and data will remain in the chosen data source (PostgreSQL, Redshift, or Snowflake). The system fetches the data at run time using the live connections on user actions.
- For the newly input data in the live connection which is inserted after successful pipeline execution for the connection, the whiz-cards will not be refreshed and the NLP updates will not happen. The corresponding pipeline needs to be run again for the refresh.
Information about the details shown in response is provided in the form of narratives. These narratives help better comprehend the numbers and data shown in the response. The narratives provide additional useful information about a metric. This additional information can be average sales, minimum sales, maximum sales, total sales, etc.
The narrative is displayed in response as shown in the following figure:
- Click the 'Narratives' icon.
- The narrative dialog shows the narrative generated around the response as shown in the following figure.
Managing custom narrative templates
Now, if you want to include specific information or if you want to add more information to such narratives, you can use the Custom Narrative Templates UI solution to create templates for custom narratives.
To access the narrative templates UI solution, from the admin console, click NLP Workbench and then click Narrative Templates.
From this page, you can clone, edit, delete, or add your own custom narrative template.
Pinboards, Cards, and Responses
- When you export an XLS file, the actionable column shows incorrect values.
- For iPad: while dragging and dropping a filter on pinboards, it displays menu options for Copy, Look Up, Translate, and New Quick.
- For iPad: while scrolling on pinboards, the filter toolbar is visible.
- When Print Chart is exported as a PDF file for the timeline query, the full data is not captured.
- In comparison response, for the ambiguous query last entity is not considered as a base entity. For example, when you ask ‘Show me Boston vs Chicago Vs Philadelphia by month’ then in the response Philadelphia is not considered as a base entity.
- For comparison response, when you enter a query for more than four entities then the last entity is not populated. For example, when you ask, ‘Show me emarun vs ofasan vs trexine vs plabenil vs arobi’ then in the response ‘arobi’ is not populated.
- For POP and Comparison, when you switch to Full Stack visual, partial is not displayed. For example, when you ask ‘Show me Boston vs Chicago by months’ then in the response, when you switch to Full Stack visual partial is not displayed.
- The ‘Insight’ option is not available for Key Driver Analysis cards on pinboards.
- The Username and the encrypted password are getting copied while creating a copy of the connection.
- Users can create a connection successfully, with a null value in Schema (For Postgresql/Redshift).
- For the newly input data in the live connection which is inserted after successful pipeline execution for the connection, the whiz-cards will not be refreshed, and the NLP updates will not happen. The corresponding pipeline needs to be run again for the refresh.
- For Amazon S3 connection: The field "Path inside the bucket" should have the path till the parent directory of the folder where the CSV files are available.
- Missing field level validations for Amazon S3, Redshift, PostgreSQL, Snowflake
- The Save button should be disabled until all the mandatory fields are filled
- Mandatory fields should be marked in Asterisk (*)
- The username and password fields should not accept space in between
- An error message should be displayed for the blank port
- Port field should accept only digits (Redshift, PostgreSQL)
- Error not displayed for null value in 'Port' field for PostgreSQL.
- When adding computations, additional computations might get loaded for created data models on UI.
- Unable to resize the columns on the Data Dictionary page to see the complete text under it.
- On the select column screen, an error should be displayed when multiple data sources have the same metric.
- When the data connection name is searched, the user is not able to select the tables and data sources from the records, and the collapse and expand button is getting disabled.
- Running multiple pipelines at the same time is not recommended, if executed, successful results are not guaranteed.
Was this article helpful?
Thank you for your feedback
Sorry! We couldn't be helpful
Thank you for your feedback
We appreciate your effort and will try to fix the article