Analyzing data with ChatGPT

Analyzing data with ChatGPT#

This guide explains how to use ChatGPT to discover valuable insights into the data you have fetched and loaded into certain destinations.

Limitations#

Using ChatGPT to analyze the data set included in a data share comes with the following limitations:

  • This feature is only available for certain destinations. Currently, these destinations are Domo, Power BI, Qlik, and Tableau.

  • Size limits apply when uploading files to ChatGPT. For up-to-date information about file uploads and relevant limitations, see the ChatGPT file uploads documentation.

Prerequisites#

Before you complete the procedure in this guide, perform all of the following actions:

  • Create one of the following destinations: Domo, Power BI, Qlik, Tableau.

  • Make sure that this destination is configured to send the data you want to analyze, for example by selecting the right dimensions and metrics, or by adding filters.

  • Create a ChatGPT account and make sure that you are logged in to ChatGPT.

Analyzing your data using ChatGPT#

To analyze the data that you have chosen to send to the selected destination using ChatGPT, follow these steps:

  1. Go to the Destinations page.

  2. Click the destination that contains the data you want to analyze.

  3. Choose one of the following options:

    • To download your data set in CSV format and then upload it to ChatGPT, click image1 Download data set.

    • To copy your data set and paste it directly into ChatGPT, click image2 Copy data set.

  4. Click Open ChatGPT image3, then log in or sign up.

As a result, you can now ask ChatGPT questions about the data set that you have uploaded or copied and pasted into ChatGPT. The following section of this guide contains some examples of the type of questions you can ask ChatGPT in order to analyze your data.

Warning

Please note that using this feature involves sending your data outside Adverity. Your data will be processed by ChatGPT. For more information about how ChatGPT processes your data, see the ChatGPT data controls documentation.

Example prompts#

You can ask ChatGPT any question about your data set to gain valuable insights into the information your data represents, the quality of your data set, trends within your data, and predictions based on your data. The list below provides some example questions to help you get started with your analysis.

To get an overview of your data:

Can you provide summary statistics for the data set?

To investigate the quality of your data:

Are there any missing values in the data set? If so, how common are they?

Do any columns contain incorrect or inconsistent data?

Are there any duplicate entries in the data set?

To analyze your data:

Can you visualize the distribution of key variables (e.g., sales, leads, website traffic)?

Are there any notable trends or patterns in the Clicks data over time?

Can you identify any correlations between different marketing channels or strategies?

Can you tell me what most people typed in before they landed on my website?

To assess campaign performance:

How did each marketing campaign perform in terms of metrics like conversion rate, ROI, and customer acquisition cost?

Are there specific campaigns or channels that consistently outperform others?

Can you identify any seasonal or cyclical patterns in campaign performance?

To make predictions based on your data:

Can you build a predictive model based on this data to forecast future sales or customer behavior?

What factors have the most significant impact on future sales or customer acquisition?

Contents