AI tools can be very good at analysing, manipulating and graphing data, but it doesn't seem straight-forward.
The most obvious solutions are to use tools integrated into the spreadsheets themselves e.g. Copilot within Excel and Gemini withing Google Sheets. Both require a subscription. Copilot might be available only in business-type accounts.
Other LLMs do not seem very good at this yet, but there are other systems designed specifically for data analysis.
The following does not show ChatGPT working on an Excel file but it does illustrate another use and some hints regarding prompting.
Microsoft Excel Copilot
Capabilities: Integrated directly into Excel, it automates tasks like formula generation, data analysis, chart creation, and pattern detection.
Features:
Answers natural language queries (e.g., "Show sales trends for Q2").
Suggests PivotTables, filters, and visualizations.
Available to Microsoft 365 Enterprise users (requires subscription).
ChatGPT Advanced Data Analysis (formerly Code Interpreter)
Capabilities: Analyzes uploaded Excel/CSV files, generates summaries, creates graphs (e.g., bar charts, heatmaps), and manipulates data (e.g., filtering, merging columns).
Limitations: Outputs modified files as downloads; requires ChatGPT Plus subscription.
Google Sheets AI Add-ons
Examples: SheetGod, GPT for Sheets.
Capabilities: Generates formulas, automates macros, and classifies data using natural language prompts.
Specialized Tools
Akkio: No-code predictive analytics; builds AI models from spreadsheets for forecasting.
MonkeyLearn: NLP-focused analysis (e.g., sentiment analysis, text classification) within Excel/Sheets.
Tableau/Power BI: AI-driven insights (e.g., Tableau’s Explain Data, Power BI’s Q&A visualization).
Python Libraries (Pandas, Matplotlib, scikit-learn)
Capabilities: Manipulate data, train ML models, and create graphs. Often used with AI frameworks (e.g., TensorFlow) for advanced analysis.
Workflow: Jupyter notebooks or scripts automate Excel tasks (requires coding skills).
Automation Platforms (KNIME, RapidMiner):
Drag-and-drop interfaces for data workflows, integrating AI models for predictions.
Trifacta/Alteryx:
AI/ML-powered data cleaning (e.g., deduplication, formatting) for spreadsheets.
Reading Data: Most tools import Excel/CSV files.
Commenting: AI like ChatGPT or Copilot explains trends, outliers, or insights in plain language.
Graphs: Automated charts (e.g., line, scatter plots) via Excel Copilot, ChatGPT, or BI tools.
Manipulation: Filtering, sorting, merging datasets, or adding calculated columns (e.g., via Python scripts or Excel Copilot).
Data Privacy: Cloud-based tools (e.g., ChatGPT) may pose risks for sensitive data.
Accuracy: AI may misinterpret context or produce errors; human validation is critical.
Formatting Requirements: Some tools require structured data or specific file formats.
Automated Reporting: Generate monthly sales dashboards.
Predictive Analytics: Forecast inventory demand from historical data.
Natural Language Queries: Ask, "Which product had the highest growth in 2023?"
For Non-Technical Users: Excel Copilot or ChatGPT for quick analysis.
For Advanced Analytics: Python/Pandas or tools like Akkio.
Enterprise Needs: Power BI/Tableau with AI integrations.
Let me know if you’d like guidance on specific tools or workflows!
Yes, there are several AI systems besides Perplexity that can read Excel spreadsheets, comment on them, produce graphs, and even manipulate the data. Here's a breakdown of some notable options:
1. Large Language Models (LLMs) with Code Interpretation Capabilities:
ChatGPT (OpenAI): With the Code Interpreter plugin (currently in beta), ChatGPT can directly analyze and manipulate Excel files. You can upload a spreadsheet, ask it to perform calculations, create visualizations, and even generate new versions of the file with modifications. It's excellent at understanding natural language instructions and translating them into data operations.
Google Gemini (formerly Bard): Gemini (once fully integrated with Google Sheets and other Workspace apps) is expected to have similar capabilities to ChatGPT with Code Interpreter. The advantage here would be seamless integration with Google's ecosystem.
Claude (Anthropic): Claude is another powerful LLM known for its strong reasoning abilities. While it may not have direct plugin support for Excel manipulation, you can often provide it with the data in a structured format (e.g., CSV) and ask it to analyze and generate insights.
2. Dedicated AI-Powered Data Analysis Platforms:
Tableau CRM (formerly Einstein Analytics): Tableau CRM is a robust platform specifically designed for data analysis and visualization using AI. It connects to various data sources (including Excel), automatically identifies patterns and trends, and provides AI-driven recommendations. It's a more enterprise-focused solution.
ThoughtSpot: Similar to Tableau CRM, ThoughtSpot focuses on search-driven analytics. You can ask questions in natural language, and ThoughtSpot will generate relevant visualizations and insights from your data. It also has AI capabilities for automated analysis.
DataRobot: DataRobot is an automated machine learning platform that can ingest Excel data and automatically build predictive models. While it's primarily focused on predictive analytics, it can also provide insights into the data and generate visualizations.
3. Python Libraries with AI Integration:
Pandas + Scikit-learn/TensorFlow/PyTorch: If you're comfortable with Python, you can use the Pandas library to read and manipulate Excel data. Then, you can leverage AI/ML libraries like Scikit-learn (for general ML tasks), TensorFlow, or PyTorch (for deep learning) to perform advanced analysis, build models, and generate visualizations using libraries like Matplotlib or Seaborn. This approach offers the most flexibility but requires programming expertise.
Here's a summary of the options:
ChatGPT (with Code Interpreter)
Reads, comments, manipulates, graphs
Very easy (natural language interface)
No programming required
Google Gemini (with similar integration)
Reads, comments, manipulates, graphs (expected)
Very easy (natural language interface)
No programming required
Claude
Analyzes and comments on data (needs structured input), can guide graph creation.
Medium (requires some understanding of data structure)
No programming required
Tableau CRM
Reads, comments, manipulates, graphs, AI-driven insights
Medium (requires learning the platform)
No programming required
ThoughtSpot
Reads, comments, manipulates, graphs, search-driven analytics
Medium (requires learning the platform)
No programming required
DataRobot
Reads, builds predictive models, provides insights, generates visualizations
Medium (requires learning the platform)
No programming required
Pandas + Scikit-learn/TensorFlow/PyTorch
Reads, comments, manipulates, graphs, advanced AI/ML capabilities, custom solutions
Hard (requires programming knowledge)
Programming required
Key Considerations When Choosing an AI System:
Complexity of Analysis: For simple tasks like calculating totals or creating basic charts, ChatGPT or Gemini might be sufficient. For more complex analysis, predictive modeling, or custom visualizations, you might need a dedicated platform or Python libraries.
Ease of Use: If you're not a technical user, a natural language interface like ChatGPT or Gemini will be the easiest to use. Data analysis platforms often have a steeper learning curve.
Cost: ChatGPT and Gemini have subscription fees. Data analysis platforms can be quite expensive, especially for enterprise-level features. Python libraries are free to use, but you'll need to invest time in learning how to program.
Data Privacy and Security: Be mindful of where you upload your Excel data, especially if it contains sensitive information. Review the privacy policies of the AI systems you're considering.
To summarize, a lot of AI systems are available that can work with Excel spreadsheets to help you get different results. I hope this helps!