Rooftop's Domain Knowledge
When creating business data science solutions, we apply the latest domain knowledge within the industry and functional area to which the problem relates. It sets the solution apart from the competitors and gives the customer's solution 'the edge'.
The key to a competitive solution starts by understanding the customer's unique business challenge and the problem to be resolved. We take a methodological approach and follow Rooftop's Predictive Analytics Model, and starts out with the latest domain knowledge relevant to the business problem. This can be industry specific as well as function specific, as for example churners, sales forecast, wind energy production.
When succeeding in developing state-of-the-art solutions, it is because the business challenge is sufficiently understood and insight on how the solution is intended to be applied. With this starting point, it is possible to identify and link industry- and function-specific domain knowledge with the economic and statistical domain knowledge. Together, it creates the foundation for developing data science algorithms resulting in competitive business data science solutions, that give competitors rear wheels.
Rooftop's Data Science Skills
Within the umbrella of data science, Rooftop Analytics offers the following data science skills.
ML og AI competences
- Machine learning; Supervised learning, Unsupervised learning, Reinforcement learning
- Computer vision with extended skills in OpenCV and Scikit-image
- Natural Language processing (NLP) with extended skills in BERT, GPT-3, and Scapy
- Python with extended skills in Tensorflow and Pytorch
- Python and R visualization packages with extended skills in ggplot and plotly
- Microsoft Power BI
Cloud services applications from
- Microsoft Azure
- Amazon Web Services
- Google Cloud Platform
Rooftop's Domain Competences
We believe that it requires strong domain knowledge in business economics and statistics to identify the most appropriate data science techniques when creating the best data intelligence solutions. The business problem and the context in which it resides, must be well understood and carefully disentangled before the appropriate machine learning algorithms can be developed.
Economics and statistical domain knowledge being applied
- Time series analysis (short and long run)
- Causal identification and estimation
- Market price formation in commodity, forex, financial, and retail markets
- Price elasticity and substitution effects
- Risk and uncertainty assessment and estimation
- Supply and demand optimization