I wanted to try analyzing Google Analytics data using ChatGPT-4. This is very basic stuff, but wanted to document the first tentative awkward steps in the process.

1. The GA4 Data and the ChatGPT-4 Analysis side by side

Below is the GA4 data from Looker Studio.

Below is the analysis of this data (the top 20 from each table at least) from ChatGPT-4:

Here is a summary of five of the most significant data points from the three tables you provided:

1. “Multiple DALL-E 2 images created from the same prompt – Michael Howe-Ely” experienced a remarkable 532.96% increase in total users, indicating a surge in interest in that specific topic.
2. India saw a substantial increase of 61.73% in total users, making it the second most popular country for your website. This highlights the growing interest in your content among Indian users.
3. Direct traffic to your website increased by 36.96%, suggesting that more users are accessing your site directly, which may indicate higher brand recognition, increased marketing effectiveness, or improved user retention.
4. Bing traffic experienced a 33.33% increase in total users, indicating that your website is gaining more visibility on Bing’s search engine and suggesting potential for further growth in user acquisition from this source.
5. Google, the primary source of users for your website, saw a slight increase of 1.60% in total users. While Google remains the top source of traffic, it’s essential to maintain and improve visibility on this search engine to continue attracting new users.

By understanding these key data points, you can focus your efforts on further engaging the growing user segments, optimizing content for search engines, and exploring additional marketing channels to reach a broader audience.

To liven things up, here’s a cute little robot analyzing a spreadsheet!

2. Steps in Analyzing Google Analytics Data using GhatGPT-4

The first thing I did was create a Looker Studio report with three separate tables. The dimensions were Page Title, Country and User Source and the metric for all three was Total Users. This is using GA4 data.

I then exported each table as a CSV. Then I opened each CSV using Notepad++, as shown below.

I copied each of the tables into Midjourney, and asked it to analyze the data.

ChatGPT-4 then provided some commentary on the data.

After it had provided data about the top 20 entries for each table, I then asked it look through the three tables and provide me with five of what it considered to be the most important points about the data.

A robot who may or may not be ChatGPT-4 (it’s not) carefully analyzing the GA4 data inside my computer.

3. Current limitations using ChatGPT to analyze data

I asked ChatGPT version 4 what amount of table data it could handle in the chat interface. It replied with the following:

“As a text-based AI model, I can handle a limited amount of data effectively. While there is no strict limitation on the number of columns or rows, it’s generally better to provide a smaller, representative sample of your data to ensure that the analysis is accurate and useful.

A good starting point could be providing data with around 5-10 columns and 10-20 rows. This allows enough information to give you some insights without overwhelming the model. However, the actual amount of data that works best may vary depending on the complexity of your dataset and the type of analysis you require.

Remember that if you need to analyze large datasets, specialized tools and software like Excel, Google Sheets, or pandas in Python are more appropriate solutions.”

I’m trying to figure out how to use BigQuery to connect large amounts of Google Analytics data to ChatGPT that I’ll hopefully write another blog post about. In the meantime I hoped you like this blog post about analyzing Google Analytics data using ChatGPT-4!

My Other Blog Posts