CodeSquire: The AI Assistant for Data Science
Recently in my spare time, I’ve been working with Daniiar Abdiev to create an AI coding assistant tailored to data scientists. Dan made a blog post about why this tool is needed not long ago. We’ve started by building a chrome extension that brings OpenAI’s Codex code completions to the browser. This allows it to work in places familiar to date scientists such as Google Colab, Big Query and Jupyter Labs. So far we have Google Colab and Jupyter Labs support.
This weekend I have been working away to add support for Big Query, another tool familiar to most data scientists. As seen in the image, just by writing what data you would like to retrieve, CodeSquire can suggest the SQL required. In the future, we will add support for better data comprehension so that CodeSquire knows about your column and table names.
Big Query support is now finished and will be added to the next CodeSquire Chrome extension release in the coming week. If you would like to try it out or try CodeSquire with Google Colab or Jupyter Labs, come start a free trial on our website. We give a one-month free trial with no credit card required.
Future Plans
We’ve created a chrome extension that connects data science and ML-related browser IDEs to OpenAI’s Codex code completions, but this is just the start. Our roadmap for the future includes adding data comprehension (e.g. your table and column names), automatic code transformations (e.g. pandas to pyspark), and much more in order to build a truly intelligent assistant for data scientists.
