DataWizards was created as a “spin-off” of the web development company Foreach
Since 2006 “data handling services” where a (second) core activity of Foreach.
These activities are now centralised in the company DataWizards.

What we do

DataWizards is offering datamining services

To create a competitive advantage in today’s market place large volumes of real time data have to be manipulated by means of machine learning and statistical algorithms. These methods can be used for either exploratory or predictive analytics. Exploratory analytics allows a client to answer for example the question: ”What do my clients really like about my service? “ . To predict how a client will react to the introduction or launch of a new product/service is a typical example of what predictive analytics can do. Another example is fraud detection. Using machine learning techniques, one could detect fraud in real time before it actually occurs. We are unique because we offer our clients’ hands-on training and flexibility during the project phase.


What is datamining? Datamining is the analysis of large data to extract previously unknown interesting patterns. These patterns could be groups of data records (cluster analysis), unusual records (anomaly detection) and dependencies (association rule mining).


Our expertise ranges from: segmentation, profiling of clients, churn analysis, regression analysis, CHARD, CART (Classification and regression tree), Naïve Bayes, SVM(Support Vector Machine), NN (Neural network), Structural equation modelling, ANOVA, Factor analysis, Principal component, advance Bayesian statistics (this includes particle filtering algorithms) among others.


We pride ourselves in the use of state of the art software which includes SAS (All modules), SPSS, R, MATLAB, RapidMiner, WinBUGS, Microsoft,Tableau, IBM, Prognoz, Actuate, SAP, MicoStrategy, QlikTech. Since Microsoft Excel is a popular data analytics tool among our clients, we can also provide our modelling and analytic service in Excel.


DataWizards starts with the formulation of a hypothesis


Sustainability and environmental care… one of the major points of the past and future decades, without a doubt, also for energy companies. Since a couple of years we are working for one of the largest energy companies in the world since a couple of years and recently we built an online tool that helps reduce carbon dioxide emissions in a very direct way..

The transportation sector is one of the major malefactors regarding the emission of CO2, and a company like our client is heavily involved in container transport and logistics. So to reduce the impact on the environment as much as possible it’s a good idea to focus on the loading of the containers: well stacked containers can hold more items which results in less transport and is both cheaper and greener.

Our client noticed that the trucks are rarely complete or fully charged when customers place an order. They already had an optimization process for this, but it was not mandatory and usually the containers are owned by the end customer anyway. Additionally they were trying to improve the load distribution using a very complex spreadsheet that was not working as they expected (not to mention not very easy to use and even a bit boring).

So, our main challenge besides getting the end customer involved in this process was to deliver all feedback contained in that nifty spreadsheet but in a more visual, intuitive and easy way without losing accuracy in calculations. So after quite some analysis of the spreadsheet, we could translate the complexity into a 3 step wizard delivered as a Flash application. A lot of focus was put on the user interface: making it both visually attractive and efficient in use.

Working with the new tool is very simple and requires almost no time or effort from the customer (and it’s a lot more fun to look at). The aim is well-loaded containers and if that can be achieved by ordering more of the same (or different) products at the same moment… that is what the application will suggest.

It’s a lot of fun - and satisfying - to use a technology like Flash to build this type of solution, but at least equally important is the call-to-action. That’s why our client has decided to make the optimization process part of the contractual agreement with its customers as of now: more users of the tool, better results. And even though it’s not going to stop a big oil leak in the Gulf of Mexico, I do think it is this type of commitment we need for the future of our big blue ball.

read more

Formulation of a hypothesis

Our research starts with the formulation of a hypothesis with our clients depending on the business needs. DataWizards has the expertise to extract the right data from the database before commencing analysis. An example of a recent project involves market segmentation for a famous brand in the US. The data to be analysed involves 40 variables with 70,000 observations.

Using a combination of statistical and machine learning techniques, such as principal component analysis (used for multi-dimensional datasets) and clustering (K-means), we were able to reduce the dimensions before partitioning into segments.

To predict the response of each client in the segment, we partitioned the dataset into training and validation data. By training the data using different leaners, we tested the algorithm using the validated data. We also used the ROC (Receivers Operating Characteristics) curve to determine the strength of the predictions before choosing the best algorithm.

read more


Questions for DataWizards? Contact us!

DataWIzards bvba
Frankrijklei 98 - Bus 304
2000 Antwerp - BELGIUM
+32 (0)3 290 79 11
+32 (0)3 290 79 11
BTW: BE 0 541 848 532