We aim and hit spot-on with our data intelligence solutions, because we see and understand business opportunities, and are able to translate it into solutions by applying the data and robust and trustworthy predictions.
When working with data science, an incredible number of features and large amounts of data are involved. On that basis, we apply machine learning to develop algorithms that predict very accurate. But the results can not be explained or interpreted. Therefore, results from the use of machine learning are considered as a ‘black box’.
At Rooftop, we go a step further. We use our financial and statistical competencies to analyze the problem. We break down and identify the areas which can be explained using economic and statistical theory, and we incorporate this into the algorithms. This makes the prediction more robust, because the algorithm achieves the ability to make more accurate predictions, even if environmental factors change, that cannot be completely decoded in the data base. This is business data science.
By breaking down the problem into sub-problems, we gain an insight, that has great value for the company. It provides knowledge about contents and interdependencies. It simplifies complicated business challenges without losing insight, and can be used constructively in business decisions.
Not all tasks call for business data science but situations where business decisions are influenced by the markets or consumers, it makes sense - and creates value.
Most of the assignments we solve, are business economics decision problems within the following areas
We provide inspiring talks to network groups, corporate management or after-hours meetings, where we focus on the many topics and insights that lead to the transformation towards becoming a data-driven company.
We offer competent feedback and 2nd opinion on business ideas, project initiatives, and roadmaps with data intelligence initiatives. Depending on the company's level of ambition, it can be performed through dialogue-based feedback, or through a deeper evaluation of the problem, solution proposals and data availability.
We prepare and carry out workshops, based on the specific challenges involved in the work of performing data intelligence projects and implementing the solutions in the organization. We provide methods, tools, and approaches on how to initiate necessary activities; for example, identifying the company’s data maturity, prepare an optimal roadmap or project portfolio, and we uncover pitfalls particularly relevant to data intelligence initiatives in your company
We strengthen your ability to see business opportunities in your data, which brings competitive and unique products and services, as our business data science competencies are applied.
We shape your business idea in a proof of concept to verify, if use of data intelligence is feasible, and whether it can release the ambition of a cutting-edge solution.
We transform the business idea into a pilot project, to uncover and test whether the required elements of the data science are in place. Besides business data science, this can also include verification of access to data and required IT platform, as well as business engagement to shape the solution and addressing the business transformation.
We prepare a design document, which explains the outcome of applying domain knowledge, economic theory, statistical modelling and machine learning, as this is the foundation for developing the data intelligence solution; a document which explains the content of the ‘back box’.
We develop the algorithms and programming code to be incorporated into your existing IT platform. At the time it’s being installed, it has been tested and trained, the prediction has been validated. Your domain competences are comfortable with the predictions as well as the validity measures identified by the confidence intervals.
We manage the launching period, when the solution start being used. We critically evaluate the results and the data applied, to verify the robustness of the algorithm. We also define and install the monitoring of the performance of the solution.
We lead data intelligence project implementation, where we use Rooftop's Predictive Analytics Model to develop the data intelligence solution and follow our customer's preferred implementation method.
We facilitate the anchoring of the solution to ensure that workflows are adobpted and users are trained to use the solution. Embedded is to give the insight to creates user's confidence in applying the solution, and not least obtaining the benefits applying it.
Smaller enhancements. Often, in-house data scientists have a broad area of responsibility as well as project participation on the agenda, which causes enhancements not being resolved, even when business benefits are proven. Therefore, the assignments should be outsourced to Rooftop.
On-the-job training. The ML-techniques evolve faster than the in-house data scientist can manage to keep up with. Therefore, there is a need to find effective ways for continuously learning. Resolving the actual assignments while having hands-on-training by an expert is the best way to learn and develop data science knowledge.
Business data scientist support desk. Over time, more data accumulates and continuously the algorithms need to be tuned. Monitoring data, measuring the performance of algorithms, validate results (within the confidence interval) and detecting irregularities is critical. It is time consuming and can be outsourced to Rooftop.