Guided machine learning

Data Science Studio guides you in the machine learning process. It helps you select the most adapted algorithm and variables depending on your business needs.

If you are new to machine learning, Data Science Studio gets you your first results quickly. Our white box approach shows you what is happening at each step.

If you are a veteran Data Scientist, Data Science Studio lets you focus on problem-solving by eliminating repetitive tasks.

Predict Events

Automatically optimized models

Data Science Studio allows you to build predictive models that help optimize your business.

Data Science Studio automatically selects the best algorithms according to the size, complexity and shape of your data, then identifies and transforms the most important variables.

Afterwards, it provides a clear explanation of your model's performance. It also creates an event probability scorer that can be used in real-time.

  • Decision Trees
  • k-means
  • Neural Network
  • Random Forest
  • Support Vector Machine
  • Stochastic Gradient Descent

Profiling

Understand process and behaviours

Data Science Studio helps you build an out-of-the box profiling out of your data. The output of the algorithm is an analysis of the typical behaviours or processes in your underlying data, with an analysis of the most relevant factors of these processes.

The algorithm can be applied for instance to segment your user base.

Variables reshaping

Audit, Select and prepare variables

The best Data Scientist outperforms others by being the most creative in combining and transforming variables.

Data Science Studio provides a convenient tool that simplifies variables preparation. It comes with a data transformation library, and can automatically select the most appropriate ones depending on your data shape.

  • Dummification / Binification
  • Selecting
  • Scaling
  • Textual features extraction

Keep control of your model

White box and open source

Data Science Studio produces models that are completely white-box, using open source algorithms.

The models are written in Python, which you can further customize to adapt to your specific needs.