Skip to main content

AI Toolkit for Data Analysts

Artificial Intelligence (AI) has been a part of our lives for many years, often in ways we might not immediately recognize. From facial recognition on our phones to the search results we see on Google, AI is deeply integrated into everyday technology. While there is ongoing debate about AI's potential, it's crucial to understand its role as a powerful assistant when used correctly. This toolkit focuses on generative AI, which can significantly enhance productivity by providing pre-packaged actions that drive value across the university. Tools like Microsoft Copilot and ChatGPT are particularly effective in processing and generating human-like text, making them valuable assets in the data analytics space.

Struggles, ideas, and prompts on how to use Copilot

Struggle: Meeting Notes
  • Idea: If a transcript is provided, paste it into Copilot.
    • Sample Prompt: Please summarize my meeting. Can you highlight key action items. What are the next steps for [name]?
Struggle: Summarizing Analysis
  • Idea: Reword emails or condense analysis to one-pager
    • Sample prompt: Please summarize as an executive one-pager.
  • Idea: Summarize someone else’s lengthy messages.
    • Sample prompt: Please summarize [paste text or attach file].
Struggle: Coming up with solutions in a dashboard to solve a problem
  • Idea: Dashboard idea generation
    • Sample prompts: [Share problem]. Please give me 3 ideas and answer all questions for each one. Be specific as possible and if you’re using data, make sure you have an idea on how to obtain and analyze the data. Questions: What is the dashboard about? Describe some of the core visuals for this dashboard?
Struggle: Calculations or Formulas
  • Sample prompts: Write an Excel formula for [describe what you are trying to do]. Write a DAX query to [describe what you are trying to do].
Struggle: Code Does Not Work
  • Sample prompt: Why is my SQL invalid?
Struggle: Massive Spreadsheets
  • Idea: Use Copilot to analyze
    • Sample prompts: Please analyze the attached file. Identify trends.
  • Idea: Ideas for upcoming projects
    • Sample prompt: [Provide previous analysis or data]. Suggest three ideas for future analyses. 
Struggle: Code Documentation
  • Idea: Paste your code to create standard documentation. This can be added in-line to the code or elsewhere.
    • Sample prompt: Write documentation using standard practices for the code below. Include an overview, usage guidelines, author, and last updated.
Struggle: Needing Another Set of Eyes on Your Dashboard
  • Idea: Drop a screenshot of your dashboard into the prompt box and ask for feedback. 
Struggle: Creating Dashboard Descriptions for the Data Catalog
  • Idea: Drop a screenshot of your dashboard(s) into the prompt box and ask for it to generate a description.
    • Sample prompt: Write a dashboard description. It should be one paragraph. Do not include any statistics or numbers in the description. Do include the purpose of the dashboard and audience. 

 

Tips for Using Artificial Intelligence

Prompts

  • Write clear, simple, and specific prompts
  • Use proper grammar, spelling, and punctuation
  • Include contextual information
  • Test and refine prompts

Be aware of ethical concerns and biases

  • Be careful when choosing training datasets; biased training datasets can lead to biased outcomes
  • Check content with a plagiarism checker to make sure you are not stealing someone else’s work. Rephrasing outputs is one way to avoid plagiarism. Another way to is to use AI-generated content as a starting point, instead of as a final result
  • Be cautious when using AI-generated content because of potential copyright issues

Reviewing

  • AI tools can provide inaccurate or made-up information; always be sure to fact check the information it gives you
  • Review output for spelling or grammar errors
  • AI tools are a not replacement; they require you to have knowledge on the subject and the ability to validate the output
  • Questions to ask yourself when reviewing:
    • Is it accurate and contextually relevant?
    • Does it align with your stakeholders/project goals?
    • Is the tone respectful and ethical?
    • Is it actually useful?
    • Could it create downstream risk?

 

Data Viz Bot

In early fall 2025, the Data Viz team is piloting a Data Viz bot. It scans Ohio State documentation around Tableau, the data lake (RAE), and data governance. This eliminates the need to wait for responses. If you would like to participate in the pilot and offer feedback, please submit a ticket to data-visualization@osu.edu. 

Important Information