Data Analytics

How to get the most out of your data

Modelling & Validation

To create value from data requires the use of mathematical models. RiskQuest has substantial experience in developing and validating all kinds of models, ranging from simple spreadsheet calculations to deep learning algorithms. To reduce model risk, we apply extensive test procedures. Furthermore, we also apply a broader view as we believe that statistics are no substitute for judgement. This broader view should lead to model outcomes that can be translated into actionable insights.

Data collection

Data-analytics starts with the data itself. The data quality in terms of correctness and completeness directly determines to what extent business questions can be answered sufficiently. Therefore we put strong focus on the data collection process in terms of defining the data model, pre-processing, cleansing, testing (e.g. for representativeness but also for plausibility) and finally setting data governance.

RiskQuest has over 15 years of experience in collecting and preparing large data sets for data analytics purposes.

Model governance

We help our clients to manage the risks related to the use of mathematical models in answering business questions. These risks refer to the chance of unintended consequences resulting from model development, inputs or outputs.

We achieve this by establishing Model governance which is a set of activities, policies and procedures which formalize model and model risk management activities for implementation. In particular model governance identifies a set of model stakeholders (e.g. model owner, model validator) and defines their roles within the process.

One specific topic relates to the use of black box types of techniques such as deep learning. To avoid unintended results such as discrimination, we pay attention to algorithmic fairness.

Business analytics

Prior to answering the question “why did something happen ?” and “what will happen ?”, it is important to focus on the question what actually happened. Our experience in data analytics tells us that providing insight into “what happened ?”, forms a solid basis for further discussion and questioning.

To help answer the question what happened, we usually create interactive dashboards (e.g. in Python) that allow our clients to slice and dice through the results.


Applying clustering techniques to global retailor

Applying unsupervised clustering techniques to rationalise products across the globe


A data-driven proactive client strategy

By exploring the data of large network provider, we helped them develop proactive client advice


Improve sensor activity for a care institution

Use of historical data to improve sensor accuracy and prioritize signals

Hans Heintz