Skip to content

Practicus AI Documentation

Practicus AI is an end-to-end Data Science platform, which works like a spreadsheet. It works for both technical and non-technical users.

You can explore and analyze large amounts of raw data and find unique insights and patterns.

Practicus AI allows you to easily transform and prepare data, and extend with coding only if necessary.

You can create AI models using AutoML and predict the future with just a few clicks. When working with complex problems, you can share your experiments exported code with the data science team, so they can take it further without starting from scratch.

Several Practicus AI features work directly inside Spreadsheets, allowing you to bring more than 1 billion Spreadsheet users into your Data Science projects.

If needed, you can click to generate the data processing pipeline code, and go to production with ease.

Practicus AI app is forever free, and the cloud platform offers a forever free tier as well. If you need more cloud capacity, you can pay as you go, which shows as a line item in your cloud invoice. No commitments or credit card is needed.

Big Picture

big_picture

Architecture

architecture

What does Practicus AI do?

Practicus AI has 4 major elements.

1) GUI (Visual) App: Free and available for Windows, macOS and Linux. The app is the center of most functionality such as EDA (Exploratory Data Analysis), Data Preparation, AutoML, Prediction etc.

2) SDK: Free Python library (3.6 - 3.10) providing programmatic access to most of the GUI app features.

3) Spreadsheets: Certain Data Preparation and Prediction functionality can be embedded inside any spreadsheet system

4) Cloud Node: Running on AWS, optional cloud node offers advanced functionality.

How can Practicus AI help me?

Different user personas will benefit from Practicus AI differently.

The common benefit will be allowing better collaboration between both technical and non-technical users.

Data Scientists

  • Data preparation/transformation: The most time-consuming part for most data science projects.
  • Improved model discoverability: Another key issue. Practicus AI allows any user to easily search and discover models from the GUI.
  • Point & click model inference: Practicus AI allows to immediately predict without the need for model hosting.
  • Export simpler models to pure spreadsheets: Unique to Practicus AI, data scientists can export relatively simple models to Excel, Google Sheets and others.

Data Engineers

  • Easier and faster ETL: Spreadsheet interface makes data transformation faster and more intuitive.
  • Data engine freedom: You can easily switch between Spark, Pandas, DASK, RAPIDS, RAPIDS+DASK (multi GPU)
  • Code Export: You can export processing steps into pure Python with the Airflow embedding code with one-click. This allows you to take data processing code as-is or with minimal change and immediately go to production.

Business Professionals

  • Explore various data sources to analyze raw or structured data
  • Build AI models with one-click and predict the future
  • Easily join technical experts for complex AI projects, where your expertise is needed the most