📅 Today’s Picks

Build Data Analysis with LangChain Pandas Agent

Problem

Do you find yourself writing the same pandas correlation, groupby, and filtering code repeatedly for data exploration?

Complex, multi-step analyses often involve tedious manual calculations and comparisons, pulling data scientists away from higher-value tasks like modeling and insight generation.

Solution

LangChain Pandas DataFrame Agent lets you analyze data using natural language, eliminating repetitive code and speeding up your workflow.

Key capabilities:

• Ask complex analytical questions in plain English
• Multi-step analysis in single requests
• Get results with automatic explanations of methodology
• Select from multiple AI models based on your query complexity

Faster Type Checking with Ty’s Rust Engine

Problem

Traditional type checkers like mypy are slow on large codebases, making iteration cycles longer and development less efficient.

Solution

Ty is a Rust-based type checker that provides instant feedback on type errors.

When testing the FastAPI codebase, Ty completes type checking 9x faster than mypy.

Key benefits:

• Significantly faster than mypy/pyright on large codebases
• Auto-checks every save for immediate feedback while coding
• Real-time IDE integration for VS Code and popular editors
• Zero setup: run with uvx instantly, respects .gitignore automatically

☕️ Weekly Finds

hyperfine [Python Utils] – A command-line benchmarking tool for measuring the execution time of commands with statistical analysis across multiple runs

SurfSense [LLM] – Open Source Alternative to NotebookLM / Perplexity / Glean, connected to external sources such as search engines (Tavily, Linkup), Slack, Linear, Notion, YouTube, GitHub and more

stanza [ML] – Stanford NLP Python library for tokenization, sentence segmentation, part-of-speech tagging, named entity recognition, dependency parsing, and more

Looking for a specific tool? Explore 70+ Python tools →