Using the AI Data Analyst

Data Insights is currently in Beta. If you are interested in opting in, please submit this form.

Introduction

Leveraging Amazon QuickSight, this tool allows you to ask questions about an MC Trade dataset: Profiles, Revenue, Payments, or Event Registrations. It will provide an answer through visualizations, reports, and even a summary that can be copied to an email.

Back to Table of Contents

Asking the Question

Choose the type of data that you want the data analyst to examine:

  • Profiles: If you have a question about a specific profile or a broader question that would look at all profiles then click on Profiles. Questions about profiles will be understood if they relate to any of the fields within the Profile dataset.

For the initial launch of Data Insights, it will not support custom fields or affiliation codes.

  • Revenue: If you have a question about a specific invoice or a broader question that would look at all invoices, click Revenue. Questions about revenue will be understood if they relate to any of the fields within the Revenue dataset.
  • Payments: If you have a question about an invoice that has been paid, click Payments. Questions about payments will be understood if they relate to any of the fields within the Payments dataset. The Revenue and Payments datasets have several fields in common, but the Revenue dataset includes a lot of line item information whereas Payments includes payment information.
  • Event Registrations: If you have a question about a specific event or a broader question that would look at all events, click Event Registrations. This includes questions about event information and not just event registrations (signups). Questions for this type will be understood if they relate to any of the fields within the Event Registrations dataset.

After choosing the type of data, it may take a moment, but you should soon land on a page with a text box to enter your question and click Ask.

It will not be able to answer your question if your question involves a field that is not within the chosen dataset. So please use the Breakdown of Datasets as a guide.

Below the text box, you may see one or more suggested questions. Click on a suggested question to initiate that analysis or enter a question from scratch.

Back to Table of Contents

Reviewing the Answer

  • The panel on the left (A) will offer a summary of the answer:
  • The panel below (B) will ask [or] “Did you mean” in case you are trying to ask a different question or you see a good follow up question.
  • The other panels (C) will expand on the summary with totals, details, and visualizations (graphs) which may include potentially related data points.

Back to Table of Contents

More Options

If you move your mouse over a panel, you will see more options in the top right corner.

  / 

  • If you see a graph icon, this will allow you to choose a different type of graph.
  • If you see a lightbulb icon, this will allow you to see insights or forecasting.
  • If you see diagonal arrows, this will allow you to maximize/restore the panel.
  • The three vertical dots allow you to export to CSV and some panels also offer an option to “View explanation” which provides a breakdown of how the data analyst arrived at its conclusion.

In the top right corner, there is a button to maximize the panel:

Immediately below the text box are two arrows:

These arrows will become clickable as you interact with the screen. Interacting includes asking another question, changing the visual type, or modifying the question by editing the "interpreted as" section.

The first arrow will undo the last interaction:

The second arrow will redo the last interaction:

For example, assume that you first ask, “what are the top 5 largest invoices?” Then you change the visual type for one of the visualizations. If you click the arrow in the top left corner or the Back button in your browser, it will take you to the previous screen, but the curving arrow will undo the visual type.

With some visualizations, you can click on part of the visualization (e.g., one of the bars in a bar chart) to exclude it or to focus on it (i.e., excluding all others).

This is also true when clicking on one of the groups in a legend (see below).

Back to Table of Contents

Advanced Options

This section is for advanced users who want to make adjustments to the question.

As you review the answer to your question, you will see an "Interpreted as" section below the text box.

When you ask a question, Amazon QuickSight maps the question to your MC Trade data. If you click on the pencil, you can see that map. From there, you can click on any of the arrows beside them to make changes to the interpretation.

This is just an example. It will look different based on your question.

Clicking on the arrow next to "top 5" allows you to change the rank (e.g., top 10, top 15).

In this example, there are two fields being used: invoicenum and adjustedlineitemamount.

Clicking on the "invoicenum" arrow or "adjustedlineitemamount" arrow will offer more options:

 

  • Used as: You can click this to change how the field is being identified (dimension or measure). QuickSight automatically identifies a field as a dimension or measure based on the data type. Click here for more information.
  • Sorted by: You can click this to change how the data is being sorted in the visualizations (ascending or descending).
  • Change measure: You click this to change the field that is being used.
  • Aggregation: In this example, it is doing a sum of adjusted line item amounts, but if you need to change it from a sum to an average, click Aggregation and choose from the list of aggregates.
  • Format: You can change the format. For example, you could change it to Currency ($8,241,954.37) or to a Number - Thousands (8,241.95K).

Back to Table of Contents

Was this article helpful?
0 out of 0 found this helpful
Have more questions? Submit a request