What is Chat GPT?
Chat GPT (Generative Pre-trained Transformer) is a Natural Language Chat Bot developed by the Open AI organization. Historically, chat bots were setup with decision trees where a user would enter an input and the bot would determine the most probable pre-determined response. Chat GPT is different in that It’s been trained on massive datasets and provides detailed answers similar to how a human would respond to a question.
The questions and responses are much more detailed and useful than you would get from asking a digital assistant like Siri or Alexa. A major differentiator of Chat GPT is that it was also trained using technical datasets and can answer questions about SQL, and Tableau Calculation Language (TCL). The technical knowledge base that it has makes it much more useful than most large language models
You can sign up for a free account through the OpenAI website here:
Using Chat GPT to Write Tableau Formulas
When creating dashboards in Tableau you can ask Chat GPT as a reference guide to create formulas. If you include table and field names, it will work those into your formulas so they can quickly be copied and pasted into the Tableau formula editor. Be aware that table names may be different, and the results are not always entirely accurate but it’s a great quick reference and a big-time saver when it works.
Using Chat GPT as an Alteryx Reference Guide
Not everybody uses Tableau Prep as their data preparation and cleaning tool. One of the more popular alternatives is Alteryx. Chat GPT is well versed in the use of the Alteryx Platform, and it can guide you through setting up workflows and setting up complex formulas. In testing we found that it was great at formulas but had trouble describing which tools we should use in specific instances.
One use case that we found interesting was Chat GPT to troubleshoot the Tableau Server connector. You can copy errors at workflow runtime and paste them directly into Chat GPT for some quick tips on how you may be able to correct it.
Using Chat GPT as a Tableau SDK Reference
We see the true value of Chat GPT going beyond acting as a reference for formulas in Tableau Desktop. Because the software is trained on a large dataset composed of many different programming languages, it can be leveraged to make the otherwise impossible possible.
For example, as a non-programmer it would be near impossible to use the Tableau SDK to create custom visuals, setup custom data connections, or even setup custom authentication methods for Tableau Server. Chat GPT’s knowledge base not only helps you gain insights into how to set these types of advanced features up, but it will generate the code needed for you to execute it without having to go through months of training in computer programming.
Using Chat GPT to write SQL Queries
Even if you are using Tableau Prep to bring in your data, it can be quicker at times to import data with a SQL Query. If you are not very familiar with SQL, you can ask Chat GPT to write it on your behalf. When prompting Chat GPT to write SQL, you can even list specific field names and it will write an example query for you. This method can be a big time saver but it does help to have some familiarity with SQL ahead of time.
Using Chat GPT to Optimize SQL Queries
Another useful method of working with SQL and Chat GPT is prompting the bot to optimize a query for you. This is not always possible, but you can try copying and pasting an existing SQL query into the bot and ask if it can be optimized. Simple queries may not be able to be optimized and often occur more often when multiple tables are being joined or unioned, or when your initial query is pulling in a large dataset with many fields.
Using Chat GPT as an Integration Guide for RPA
It’s not always possible to directly connect to the data that you need with one of the built in Tableau connectors, SQL, or local .csv and .xlsx files. When that happens, a common solution is Robotic Process Automation (RPA). Bots can be setup to automatically run reports and interact with a computer the way that a human would. The data can then be sent to Tableau for analysis and visualization.
UiPath is one of the most popular RPA solutions on the market today. Here’s Chat GPT explaining how to integrate it with Tableau. UiPath is an enterprise grade solution, but if you need a similar solution there are several open-source solutions available that can also automate processes and feed data into Tableau.
Will Chat GPT be integrated into Tableau?
It’s unlikely that Chat GPT’s exact technology will be integrated into Tableau. OpenAI the developer of Chat GPT recently received a very large investment from Microsoft making it unlikely that Sales Force, the owner of Tableau would be able to take advantage of the technology. Microsoft will more than likely keep the technology for themselves and integrate it with Power BI and other components of their Power Platform such as Power Apps and Power Automate.
Sales Force does have its own suite of Artificial Intelligence tools, and has been aggressively investing in predictive analytics, and text analysis through their Einstein Suite of AI and Machine Learning tools. They will likely be incorporating a Natural Language Processing tool similar to Chat GPT in the near future across their platform.
Overall, we would love to see more applications of machine learning in the Tableau platform. The pieces are there, and with Microsoft aggressively introducing new AI insights, and anomaly detection features Tableau will need to stay ahead of them to maintain their lead in market share.
Could Chat GPT Perform Data Integration and Data Preparation?
Chat GPT is already capable of generating Tableau formulas, SQL Queries, and describe how to connect Tableau to other sources. We were curious if it could also prepare data using Tableau Prep or Alteryx.
We found that Chat GPT could accurately help us create a workflow when we got stuck on a spot or did not know which tool to use, but when asking similar questions regarding Alteryx Chat GPT would more often return inaccurate information or would get tools confused such as a Join Tool vs. a Union Tool.
This will likely improve in the future as the dataset that Chat GPT is trained on learns from users, or the program is upgraded to take advantage of a larger training dataset.
Would Chat GPT be Useful for Building Power Automate Flows?
Sales Force offers its own process automation solution called Sales Force Flow which is great for creating low code workflows, especially if you are integrating them within the Sales Force Platform.
As a more approachable solution with a wide range of applications, we really like Microsoft’s Power Automate Flows for their ease of use and ability to quickly deploy as long as you have a Microsoft 365 Enterprise Account. Power Automate Flows are getting even easier to create now that Chat GPT style technology is being used to convert a natural language text description of a process into an automation.
Click here to learn more about Microsoft’s Describe it to Design it function:
Chat GPT and similar solutions will bring the amount of time to develop RPA from hours to minutes. The bigger challenge though may be finding people that can eloquently articulate the steps of a process in a way that a chat bot could decipher it. Have you ever been on a Microsoft Teams call sharing your screen while trying to explain which button they need to push? The one to the left! No, the one to the other left!
How to use Chat GPT to Generate Sample Data Sets
Because we do a lot of consulting and create proof of concept applications, automations, and reports we are constantly creating example data sets. Chat GPT makes content creation much faster.
As an example, we asked Chat GPT to generate a 20-row dataset for sales and sales quantity by product for a pet store. The cool thing about using Chat GPT for this is that the products are contextually appropriate for a pet store and can be generated in seconds with zero thought and without opening an Excel spreadsheet.
There are limitations on the amount of data that Chat GPT will generate for you. When prompting the software to generate a large set of data it would not complete the request and would often stall out part of the way through. If you need lots of made-up data we recommend solutions such as Mockaroo – Random Data Generator and API Mocking Tool | JSON / CSV / SQL / Excel or one of the many other free tools available. Chat GPT is nice because we typically already have it open and don’t need large amounts of data for proof-of-concept work.
What are the Limitations of Chat GPT?
Chat GPT is already changing the way that people are interacting with computers. It’s upending the way that people create and search for content. Chat GPT makes life easier for business analysts and business professionals developing Tableau Dashboards. As the program gets better it will likely be integrated into more software solutions and act as a trusted resource.
Chat GPT is Only as Good as the Dataset it’s Trained On
The biggest challenge with Chat GPT is that the information it provides is not completely accurate. The answers that it provides are less accurate as the software solution or topic that you are asking about is more niche. We have found it to be a great resource for Python, Power BI, and Tableau as there is a huge amount of publicly available information on the internet. The answers become less reliable with software packages less widely used such as Alteryx.
Tableau Dashboards are Visual and Very End User Specific
A lot of times people don’t know what they want until they see it. They may have a rough idea that they’re tired of looking at Excel spreadsheets and would like to see more dynamic KPI’s but don’t appreciate the amount of effort that goes into making a Tableau dashboard look nice and neat. You could throw a lot of information at people, and provide the deepest level of data analysis possible, but editing takes a long time to present concise information that is easy for less technical people to understand.
We asked OpenAI Dall-E 2 to create a Tableau Dashboard
A second algorithm published by OpenAI, the developers of Chat GPT is Dall-E 2. The algorithm takes natural language and generates pictures. It’s incredibly interesting technology as it will generate life like photographs of people, paintings and inanimate objects.
Without having access to any specific data, we asked OpenAI’s Dall-E to create an image of a Tableau Dashboard. Below is an example of an AI generated Tableau Dashboard for the Automotive Retail Industry.
As you can see it’s super rough at the moment. To be fair, Dall-E is an algorithm designed and trained to re-create photographs and does a very good job at it. There will likely be a point in the future where similar technology is deployed in Tableau as an upgrade to the Show Me feature, where Tableau suggests a potential visualization based on the fields you select.
If you’d like to try out OpenAI Dall-E image generation, check it out here:
What is the Future of Chat GPT and AI for Tableau?
Chat GPT is a tipping point for Artificial Intelligence. The technology itself has been developed over many years and is not new. The public interest in the technology and usefulness of a generally trained chat bot is new however, and it’s likely that it will be Interated into thousands of products over the next couple of years. Almost every widely used software could use ChatGPT to for an improved customer experience when a user can ask a question and receive an appropriate response.
The technology is incredibly useful as a reference to help people develop solutions faster. As the technology evolves, we may soon see a day when natural language is converted into actual steps taken by a computer. One such solution in development is ACT-1 developed by Adept.AI whose aim is to essentially let a computer automate itself after being prompted.
The pace of change is only going to speed up from here. Stay tuned to learn more about groundbreaking technologies and how you can use them to your advantage today!
Continue learning more about Chat GPT’s promise to revolutionize the future of business intelligence and how you can start taking advantage of it today regardless of whether you work in accounting, finance, or another department.