Data Compass

Business Processes

It all starts when our clients approach us with a more or less specific question or problem for which they are seeking a solution. Together, we work to gain a comprehensive understanding of the issue. To do this, we clarify and evaluate all the background context, motivations, and interests associated with the question. The more precisely we define a question, the better it can be answered using the right data. We believe it is crucial for success to have everyone involved at the table from the very beginning. The sooner all affected departments and decision-makers are included in the process, the easier and more effective the implementation will be.

Once the specific question has been defined, we move on to planning and clarifying the framework conditions of a project. To do this, we analyze the business processes to identify the technical and analytical challenges that must be overcome on the path to a solution. Understanding the interplay between business processes and the analytical concept forms the foundation.

Data Intelligence

In the next phase, we translate business- or domain-oriented questions into data-driven questions. This involves precisely determining which numbers, metrics, and data points are relevant. In some cases, the data may already be available; in others, we first need to develop concepts for data collection.

The real challenge is making very different types of data comparable with one another. In layman’s terms, one could say that data “must speak a common language.” This aspect of a data science project can be extremely time-consuming, and it often involves manual processes. That is why data intelligence is one of the crucial steps without which no reliable conclusions can be drawn. Data science requires “good” data—that is, relevant, structured, and valid data.

Predictive Analytics

Today, companies are often faced with enormous volumes of data. To analyze this big data, specialized algorithms are used or developed in-house that are capable of identifying patterns and regularities in large data sets. The analysis of historical data can, for example, be used to develop predictive models. This involves calculating the probabilities of a specific event or scenario occurring.

This allows trends to be identified early on and acted upon. Barack Obama’s election campaign became famous for this approach; his campaign team analyzed massive amounts of tweets from social media and adjusted their campaign strategy accordingly. With the help of predictive analytics, we can either evaluate specific questions and hypotheses or search for the next (meaningful) question suggested by hidden patterns and regularities in the data.

Insights Visualization

The human brain processes data in the form of images much faster and more effectively than endless rows of numbers in tabular form. Insights visualization—the visualization of data—is therefore not only important for presentations but, above all, for understanding and interpreting the information. Visualization thus becomes an essential component of any analysis.

Especially since data science is not aimed exclusively at IT experts but also finds application in management and executive leadership, the results must be presented clearly and comprehensibly. Usability and information design are key to ensuring that data science can become an integral part of everyday business practice.