You’ve likely heard the buzz about AI tools like ChatGPT and Google Bard.
After all, ChatGPT took the world by storm in late 2022, becoming the fastest product of all-time to reach one million users… in just 5 days!
While it’s clear these chat-based Large Language Model (LLM) tools are popular, it might not be as obvious how you can apply them in your workflow as an Analytics pro. That’s what we’ll be talking about today.
Why Data Pros Need To Learn To Use AI Tools
The world is changing, fast.
Tools like ChatGPT and Google Bard are changing the game, allowing users to complete tasks in minutes that used to take hours.
Here are just some of the reasons any data pro should consider using AI:
- Improve your performance & efficiency
- Automate routine, low-value tasks
- Spend more time on the higher value activities
- Use AI as an easy way to check your work
- Leverage AI tools to learn and get questions answered quickly
- Stay on the cutting edge and competitive in your field
Will AI Tools Replace The Need For Data Analyst And Data Scientist Jobs?
This question gets a lot of people pretty fired up… especially folks in these roles or studying to enter the field. Makes sense, it’s a scary concept.
But fears of Analytics pros losing their jobs to AI are totally overblown.
Companies will always need the special sauce that truly great Analysts bring to the table:
- Strategic thinking / problem solving
- Communication skills
- Technical proficiency
AI tools can help with the technical gaps, but when it comes to thinking strategically about a business, deciding where to focus, advocating for and driving change within an organization, AI isn’t showing much promise.
For the foreseeable future, any solid Analyst with good strategic thinking, problem solving, and communication skills is probably safe from AI taking their job.
While AI likely won’t take your job, another Analyst who uses AI might.
Imagine trying to be an Analyst today, and never having learned how to use spreadsheets or the internet. You’d be at such a disadvantage relative to your competition. We’ll be saying the same thing about AI tools very soon.
To make sure you continue to be seen as a top talent in data, you’ll need to work on the uniquely human skills, build a solid technical foundation, and learn to use cutting edge technology like AI. If you can do that, you’ll be truly unstoppable in your career.
Warning: Common Pitfalls of Using LLM Tools
Before we get into specific analytics use case for AI, it’s worth noting some of the things you need to watch out for while you’re using them.
- Don’t get yourself in trouble by sharing sensitive company data. Your information isn’t private once you share it with an LLM. A number of companies have already banned employees from using these tools. Any employer would be upset if they found out you compromised sensitive information externally. Don’t get yourself fired.
- LLMs are known to “hallucinate”. This is a nice term for saying they make stuff up. They do it all the time. The problem is, these tools sound just as confident when hallucinating as when they are giving you accurate information.
- Sometimes you’ll get an answer to your question, but the solution you receive might not be the optimal one.
- Today’s LLM tools are very broad and they might lack specific domain expertise or deep knowledge of a specific area where you need them to perform.
- LLMs are NOT capable of common sense, and cannot (at least today) replicate human judgment. These models are trained on the open internet. A lot of the information out there is bad, and the models aren’t good at making that distinction.
The general rules of thumb here are to use sound judgment about what you should and shouldn’t be doing with these tools, and think critically about the responses you’re getting back.
If you can do that, and remain aware of the current limitations of LLMs, then they can be really valuable tools that will help you speed things up and automate some of the more mundane tasks in the analytics workflow.
It’s also important to note that these tools are improving very fast. Many of the limitations we are discussing today are problems that may be solved in the near future. Keep these pitfalls in mind, but considering everything related to AI tools a moving target.
Best Practices For Using AI Tools For Analytics
One of the quickest ways to be more effective with these tools is to learn how to write better prompts.
Here are some of our Prompt Engineering tips to help you get the most out of LLMs like ChatGPT and Google Bard:
- Be clear and specific. The more detailed information you give the AI, the more likely you are to get a high value response.
2. Provide context. The more information you can share about your situation and the problem you are trying to solve, the better.
3. Establish roles. Telling the AI tool who it should impersonate, and who the response is for can go a long way toward shaping a relevant response.
4. Set the tone. Try prescribing things like how technical, how detailed, how formal, and how long you want the response to be.
These are good rules of thumb, but don’t overthink them at the expense of speed. It’s easy to scrap your conversation and start fresh if you don’t like where things are headed. So aim for good progress here and don’t worry about perfection when it comes to prompt writing.
Best ChatGPT Use Cases For Data Pros
There are TONS of ways you can use these AI tools in your workflow.
Some of our favorite use cases include:
- Generating code, queries, or formulas
- Troubleshooting or debugging your code
- Adding human readable comments to your code
- Performance optimizing code, queries, or formulas
- Automating manual tasks
- Providing data visualization tips
- Generating data samples
- Explaining a technical concept
- Generating step-by-step tutorials
There are plenty more, but these are the ones that keep coming up.
Next we’ll go into a bit more detail about how you can use AI with tools like Excel, Google Sheets, Power BI, SQL, and Python.
Excel Use Cases For ChatGPT
Here’s an example showing how you can use ChatGPT to generate an Excel formula from scratch:
Here are some of our favorite Excel use cases:
- Explaining how an Excel formula works or what it’s doing
- Generating formulas from scratch (illustrated above)
- Creating DAX or M code
- Generating VBA scripts or automation (video below)
- Troubleshooting errors in formulas or code
- Data prep & exploratory data analysis
- Generating sample data
The video below gives a more detailed walkthrough, where Chris shows us how you can use ChatGPT to automate tasks with VBA:
Google Sheets Use Cases For ChatGPT
Here’s an example of how you can use ChatGPT to troubleshoot formula errors in Google Sheets.
Some of our favorite ChatGPT use cases for Google Sheets are:
- Explaining formulas
- Troubleshooting errors (pictured above)
- Generating formulas from scratch
- Applying formatting
- Adapting Excel tools for Google Sheets
- Writing Regex patterns
- Coding apps scripts (see video below)
Below is a great video from Enrique showing us how you can use ChatGPT to code an apps script (Google Sheets’ version of macros).
Fun fact: this is the first apps script Enrique ever wrote, so you can see it from the perspective of a non-expert.
Power BI Use Cases For ChatGPT
Here’s an example showing how ChatGPT can generate DAX measures from a fairly simple description:
There are lots of good use ChatGPT use cases for Power BI. Here’s a list of some of our favorites:
- Connecting to data sources (video below)
- Understanding how to get started
- Explaining specific Power BI concepts
- Creating DAX calculations
- Explaining what code is doing
- Generating measures
- Troubleshooting errors
- Creating visuals and dashboards
In this video, Aaron walks through one of our use cases, using ChatGPT to figure out how to connect Power BI to a specific data source.
SQL Use Cases For ChatGPT
Here’s an example showing how you can use ChatGPT to add human readable comments to our SQL queries.
Your coworkers and future self will thank you.
These use cases for pairing ChatGPT with SQL can be hit or miss. Some of them are great today, and others still need some work. I think they will get much better in a short amount of time.
- Explaining a SQL concept
- Describing what a SQL query is doing
- Comment SQL code (pictured above)
- Debugging errors in your query
- Generating SQL queries from scratch (video below)
- Performance optimizing your query
Here’s a video where I show you how to create SQL queries from scratch using tools like ChatGPT and Google Bard:
Python Use Cases For ChatGPT
Here’s an example showing how you can use ChatGPT to performance-optimize your Python code.
Some of the other use cases you might want to try with Python are:
- Explaining Python code
- Troubleshooting errors in your code
- Performance-optimizing your code (pictured above)
- Generating Python code from scratch (video below)
- Researching libraries
- Web scraping
- Interpreting Machine Learning models
- Jupyter Notebooks plugin (not free)
Here’s a more detailed walkthrough where Chris walks us through how to create Python code from scratch using ChatGPT.
New Free Course: ChatGPT For Data Analytics
If you enjoyed reading this one, and want to learn more, Maven Analytics just launched a brand new course: ChatGPT for Data Analytics.
In this course, we’ll introduce you to the world of deep learning and generative AI, explore the rapid rise of large language models like ChatGPT and Google Bard, and get you up and running with free tools that will take your skills to the next level.
These are just some of the topics we’ll be covering…
- Why AI for Data Analytics?
- Intro to AI, LLM’s, & ChatGPT
- Prompt Engineering
- ChatGPT for Excel
- ChatGPT for Google Sheets
- ChatGPT for Power BI
- ChatGPT for SQL
- ChatGPT for Python
- …and more!
This course is beginner-friendly, and designed for anyone who wants to leverage modern technology to work more efficiently, and make smarter, data-driven decisions.