How we work?

Learn, in what way we work.


Simply, each of our projects is proceed to optimize its positive business impact. We work using a methodology close to CRISP DM.

  • We sign a Non-Disclosure Agreement (NDA).
  • We talk about your business and goals for the project — we need to know how your company works to achieve the best results.
  • We sign an agreement for a small project, a feasibility study or a reconnaisance project.
  • You deliver a small sample af data and / or their description (if needed then in some of projects the data can be anonymised).
  • We discuss the data and our conclusions from their analysis (remote or in person meeting).
  • Jointly decision about the next steps.
  • Definition of a clear business goal for the project.
  • Definition of scope and duration of the project.
  • Signing of an adequate agreement.
  • We start to work contacting you if needed.
  • A control point (remote or in person meeting):
    • delivering of intermediate results of work
    • talk
    • in case of typical data exploration projest we discuss with your expert about relationships, patterns, and conclusions discovered by us in the data
  • Finishing of the project and implementation of the results into the business.
  • Helping in implementation — if needed.

Projects can proceed other ways:

  • For small and big projects the workflow can be different.
  • Duration of particular stages differs depending of the size of the project.
  • For small projects a preliminary project or feasibility study are not needed.
  • For building of predictive and scoring models we do not proceed with a preliminary project — it is not needed because these kind of analytical project are well standardized.


  • We deliver conclusion from the analysis plus reasoning for them. This is because our goal is not only discovering from data useful, previously unnknown, knowledge. We want to understand why it is and transfer this information to a client.
  • If we are building a model then we deliver it in a readible form — we do not hide it as it is sometimes done.
  • A predictive model is supplied with a report that contains information about the methods used. We never deliver a ”black-box” type of model what facilitates your internal experts to value our work.
  • We justify our methodological decisions taken during modelling.
  • Software we develop is delivered along with a suitable documentation.
  • When the software is developed in R then it is delivered with the source code.
  • We supply well described algorithms if our goal was to develop them or a client has to implement them on his own.


What methods do whe apply? Methods, that using we teach and many other — depending on the needs. Look into programs of the following courses to learn what methods we use:


We base foremost on free software what often reduces client cost of implementing our solutions comparing to using proprietary software:

We benefit from using commercial tools too. It is foremost for software development projects. Of course we work in MS Excel too.

Try again