Choice of Application for Developing Algorithms

In our work of developing algorithms, we use various technologies and software programs. We make extensive use of open source applications and cloud, and we integrate with various technologies on many platforms.

The starting point for Rooftop Analytics is always the client's unique problem and the available data when it comes identifying data model and program algorithms for predictive analytics and transcribing. We compile the program code into a software application, which is installed on the company's IT platform.

We are constantly investing in obtaining most recent knowledge on the applications and technologies on the market. To obtain the best foundation for choosing the right tool when resolving the client assignment, we use the applications and test them in practice.

Programming Algorithms in the Data Model

  • We use all types of machine learning (supervised, unsupervised, reinforcement learning).
  • We program in Python and use the TensorFlow and PyTorch libraries.
  • We also use the program language R for the predictive analytics algorithm specifically.
  • We use computer vision and natural language processing (NLP) for the data mining algorithm and development of predictors.
  • We share and document our code in GitHub or Slack.

The Predictive Analytics Software Application

  • Contains a user interface with visualization programmed in R or Python for easy access to data results.
  • Suitable for being incorporated on the client's IT platform. This being as algorithm in an automation solution for process optimization or a reporting task.
  • Generates data to be leveraged for the client's business analytical purposes.

The Software Application for Data Capture and Transcription

  • Transcripts information from handwritten documents to a data set, for data analytical purposes in companies, or research institutions.
  • Incorporates the software application into apps to leverage microtasking and crowdsourcing in the task to capture data.
  • Integrates the software application on-premise or in the cloud, optionally using API as a software component.


Are you curious about how we keep up-to-date with the latest technologies and make use of them, contact Torben Skov Dyg Johansen.

Book a meeting

read more

Our model secures we uncover the topic, are firm in our approach to develop the algorithm, and work agile.

We are a team with unique competences in data science and capable of seeing extraordinary opportunities in data.