AI Tools That Make Data Analysis Easier

AI/MCP Data Analysis Journey Day 1 - my attempt to perform an in-depth BI analysis entirely within my MCP client starts with deciding which tools I'm going to use and why... MCP CLIENT Claude Desktop - while I use other MCP clients too, I chose this one for a number of reasons: 🔸 Anthropic developed the protocol, and Claude currently is the smoothest operating experience IMO 🔸 Its visualization and artifact capabilities are currently the best IMO 🔸 It interfaces very well with Supabase, exposing the full set of Postgres SQL read/write tools AI MODELS 🔸 The native models, Opus, Claude and Haiku represent an excellent range of capabilities. However, I've also brought the Openrouter MCP tool, providing me access to 450+ additional models, so I should be all set... MCP TOOLS Beyond the set of daily app tools (gmail, Google Calendar, Slack, etc.) and OpenRouter, I think the following are going to be most critical for this journey: 🔸Supabase (28 tools) - MCP-enabled Postgres database for modeling and querying the data 🔸Desktop Commander (18) - enables operations on all my local files and can run terminal commands - giving me the ability to run Python and R within my MCP client, and access to over 30,000 specialized packages 🔸 Knowledge Graph Memory (9) - gives me persistent memory across models and also helps me overcome Claude's small context window 🔸 Obsidian MCP (12) - provides me full access to my entire Obsidian vault, with over 1,500 total notes, incl. my entire past content, other important posts, and my prompt library 🔸 Excalidraw (11) - AI drawing tool for network and data model diagrams, mermaid charts, etc. 🔸 Excel (7) - gives me access within my MCP client to the Swiss Army Knife of data 🔸 PBIXRay (14) - gives me direct access to all of my Power BI semantic models w/o even opening the files 🔸 Vegalite MCP (2) - while probably overkill, gives me the ability to render any visual in vegalite (same data viz language that powers Deneb) 🔸 Brave Browser (2) - gives me powerful web search and retrieval within my MCP client 🔸 Convert API (16) - allows me to convert directly across a wide range of file types using natural language Because MCP is a new and evolving protocol, security is very much a concern. In my installations above, I've followed the these practices: 🔸 Installed the majority of my MCP servers using Zapier and Pipedream, which provide secure API endpoints, end-to-end encryption, security monitoring and other best practices. Open AI, Anthropic and others have partnered with them to provide users easy install and secure operation 🔸 Chosen MCP servers developed by reputable companies and developers, and reviewed by Pulse MCP, which aims to fill a similar role to CRAN for R 🔸 Chosen popular servers w/ many other users (some strength in numbers) I feel confident that my choices provide a powerful, well-rounded set of capabilities for the rest of the journey. Onward! #ai #mcp #datajourney

Spreadsheets are changing rapidly. We are seeing a lot of new AI tools built into Excel that allow people to be way more productive. A few examples include: ➡️ Rosie AI - Chat GPT for Excel. I think of it as an Excel assistant for more advanced users and finance pros ➡️ Microsoft CoPilot - When I use CoPilot, I think of a tool built for the average Excel user, not the average finance Excel user. It can help, but I believe it’s being designed for the average user, not the advanced users ➡️ Other tools are in Beta, including Cursor for Excel, Shortcut, and others. I regularly hear about tools with amazing capabilities. A new generation of connected and AI-assisted spreadsheets ➡️ Equals - Used by a lot of startups, it is the spreadsheet reimagined ➡️ Row Zero - 100% cloud spreadsheet designed to handle millions of rows and manage the security of your data. It is also easy to connect databases and run SQL. ➡️ Sourcetable - A cloud spreadsheet focused on being an AI assistant, allowing you to easily connect and store millions of records in a spreadsheet interface. It includes Python and SQL ➡️ Quadratic - An AI-based spreadsheet in the cloud. You can write and edit code directly in your cells ➡️ Rows.com—I've been impressed with how easy it is to connect different data. It has gained strong traction in the marketing space. It recently added a plaid integration for banking data. ➡️ Paradigm - Spreadsheet interface with a group of agents to help you easily pull and append data and get work done. The reality is that we should be much more productive with spreadsheets than we have ever been, whether it is Excel with AI integrated or a new connected or AI-based spreadsheet. I have never been excited to see boundaries being pushed with these tools, and I think it's amazing how they will assist us in conducting analysis, modeling, and Excel work in the future. It is time to realize that the way we work in spreadsheets is changing faster than ever, and with a little bit of commitment, we can be substantially more efficient and effective than we ever thought possible. What is your favorite Excel AI agent or new spreadsheet that has you excited?

Snowflake Summit just unleashed a tidal wave of announcements! The key takeaway? Snowflake is making it easier than ever to activate ALL your data for powerful AI, faster analytics, and streamlined operations. Here are a few highlights that stand out and why you should care: 🤖 Intelligent AI, Simplified: - Snowflake Intelligence: Imagine business users conversing directly with all their data (structured & unstructured!) using natural language to get answers and take action. No code, just insights. (PuPu soon) - Cortex AI SQL: Analysts can now use familiar SQL to tackle complex AI tasks on documents, images, and more. Game-changer for productivity! (PuPu) - Data Science Agent: Automating ML workflows with natural language greatly boosts data science teams. (PrPr soon) ⚙️ Data Engineering Supercharged: - Snowflake Openflow: Effortless, managed data movement from virtually any source, crucial for feeding those AI models. (GA on AWS) - Native dbt Projects & Workspaces: Streamlined development and collaboration for building robust data pipelines. (PuPu soon) - Enhanced Apache Iceberg Support: Truly open and governed lakehouses are here, making your data more accessible and performant. (PuPu soon for key features) 📊 Analytics at Warp Speed & Scale: - Standard Warehouse – Gen2: A 2.1x performance boost for core analytics! Faster insights, better TCO. (GA) - Snowflake Semantic Views: Consistent business logic across all your tools for trustworthy analytics. (PuPu) - SnowConvert AI: Massively accelerated and de-risked migrations from legacy systems. 🚀 Platform Power & Collaboration: - Snowflake Adaptive Compute: Warehouses that just work, automatically optimizing for performance and cost. (PrPr) - Horizon Catalog Enhancements (incl. Copilot): Simplified governance and data discovery using natural language. (PrPr soon) - Expanded Marketplace (Agentic Apps, Knowledge Extensions): Easily bring in third-party AI and data to enrich your applications. Why does this matter to Snowflake customers? : ✅ Accelerating time to value ✅ Empowering more users ✅ Reducing complexity & cost ✅ Innovating with confidence Want the whole picture? I've summarized all the announcements, broken down by product category, with a clear focus on the business impact and value to you in this comprehensive blog post: 🔗 https://lnkd.in/ggB7NHHE