On Wednesday, June 27, Geert Verstraeten has animated an interactive workshop at the Data Innovation Summit, the largest conference in Data Science in our region. The conference topic is Artificial Intelligence Demystified -and in this context, Geert presented an overview of concrete and cool applications in the field of data science today – some other cool ideas and suggestions of the audience were added to the slide notes here.
On Wednesday, June 27, Kevin Heyman (data scientist at Van de Velde) and Nele Verbiest have presented their success story at the Data Innovation Summit, the largest conference in Data Science in our region. The conference topic is Artificial Intelligence Demystified – and our image-based recommender system features as a concrete success case of deep learning in practice. Cannot join? Find the slides here or read our story here: Lingerie Recommendations at Van de Velde
On Thursday, April 26, Kevin Heyman (data scientist at Van de Velde) and Nele Verbiest presented their recent success in recommending the next lingerie purchase at Van de Velde. In an exciting project, they build an image-based recommender system. Their story shows how Van de Velde succeeds in this agile data science project that delivered immediate value. From a technical point of view, we built a valuable recommender system using deep learning based on lingerie images.
Missed the presentation? Read our story here: Lingerie Recommendations at Van de Velde
Jan (with nickname Honza) joined Python Predictions as a Data Scientist in October 2017. Before joining Python Predictions, Jan has gained several years of experience in banking and a consumer protection organization, where he was working as a Data Scientist on projects involving Advanced Analytics and Business Intelligence. He is currently also finishing his masters degree in Statistics at KU Leuven.
Anthony joined Python Predictions as Data Scientist in August 2017. He holds a Master in Electromechanical Engineering and obtained an additional Master in General Management. During his master thesis and multiple internships, Anthony developed his first modeling experience. Before Python Predictions, he was mainly active in the recycling sector and public sector as a business consultant.
Elien joined Python Predictions as a Talent Manager in July 2017. She holds a master’s degree in educational science from Ghent University, and gained five years of experience in human resource management before joining.
Elien’s main responsibilities are retention of talent and recruitment of new team members. Besides this, Elien is also in charge of general office management.
Adrien joined Python Predictions as a data scientist in June 2017. He acquired his master of Management science at the Solvay Brussels School and his master of Actuarial science at University of Brussels. He developed his first modeling experience during his master thesis.
After his studies, Adrien started his career as an actuarial consultant, developing his skills in financial modeling and forecasting techniques.
Today, an estimated 8.5% of the Belgian employees are at risk of burnout. This represents a huge potential cost to organizations and a huge personal risk for employees. Together with SD Worx, a large Belgian HR service provider, we succeeded in predicting expected future absenteeism, by combining historical absenteeism with softer measures like evaluations and surveys. On March 30, we shared the main conclusions of this exiting project at the Data Innovation Summit 2017 in Brussels. And we offered our view on the four reasons why we succeeded in this challenge.
In October 2016, Sofie joined Python Predictions as an analyst. Sofie is a master in civil engineering and advanced master in water resources engineering and statistical data analysis. Before Python Predictions, Sofie was active as a project engineer in an engineering and consultancy company specialized in a vast range of water related projects.
Unfortunately, Sofie has left us in February 2018, we wish her all the best of luck in London.
Today, the 21st of April 2016, we are celebrating 10 years of Python Predictions.
Thanks to our clients, employees, partners and fans for the great journey and helping us realize our dreams!
On Monday April 11, we have demonstrated how we applied Process Mining to improve the customer experience in a crucial customer-facing process at ING Belgium. In this case we compared traditional (six sigma-style) approaches with more modern techniques to help ING Belgium serve its clients better. We illustrated the benefits, milestones, requirements and pitfalls we encountered. Presentation held on the INFORMS conference on Business Analytics and Operations Research in Orlando (USA).
Klaas joined Python Predictions in April 2016 and combines a master in bio-engineering with an advanced master in economics. After working abroad in the non-profit sector, he has built experience in advanced analytics and predictive modeling for several years within the financial sector and in fundraising.
On January 18th 2016, Guillaume Marion joined Python Predictions as an analyst. He acquired his master of Business Engineering at the University of Brussels and developed his first modeling experience during his master thesis. After his studies, Guillaume was active as a researcher in areas such as segmentation and preference forecasting.
At Python Predictions, Guillaume will primarily work in the telecommunications market. Welcome Guillaume!
We are pleased to announce that our next training for managers is scheduled on Thursday December 3rd 2015. In this intuitive, non-technical training, we will cover the essence of predictive analytics. Participants will learn how to define projects, how to manage projects and how to communicate with their analysts.
On Oct 2nd 2015, together with ING Belgium, we have presented ING’s current developments and usage of predictive analytics and data science at a conference entitled ‘Etes-vous prêts à accueillir le Big Data?” at IESEG School of Management, Paris. About ten years ago, ING Belgium started its first activities in Predictive Analytics for Marketing at ING Belgium. Since then, the bank has succeeded in transforming the first sporadic successes to a successful business activity in production. In more recent years, the data science team emerged to focus on new applications, new technology and new data.
On August 3rd 2015, Jan Muermans joined our team as an analyst. He has a master in Business Engineering from the University of Leuven, where he specialised in the fields of statistics and optimisation.
Jan started his career in FMCG, where he obtained strong business understanding. At Python Predictions he wants to combine his business sense with his knowledge of statistics in the fascinating world of predictive analytics. We are looking forward to working together with Jan!
In this first workshop Predictive Analytics specifically designed for managers, participants have learned how to define projects, how to manage projects and how to communicate with their analysts.
The workshop was hosted by the Brussels Data Science Community on Friday June 26th. This jargon-free workshop focused on the absolute essence of predictive analytics from a manager’s point of view. For practical details and future opportunities, please visit the website of the Brussels Data Science Community.
While the application of predictive risk analytics is very well established for many financial institutions, many other organizations today do not fully grasp or exploit the potential data residing in the internal customer databases to predict future credit risk for new and existing clients.
On May 21st, Geert Verstraeten has introduced the basics of predictive analytics on an event organised by CFO Magazine on the topic of working capital management. He also presented the story of our ten year collaboration with Unigro, omnichannel distant seller in the Otto Group. Find more details about this story on the site of SAS Institute.
Predictive Analytics has a tremendous impact on many domains such as for instance marketing, risk and operations. One of the major new interest fields is Human Resource (HR) management. Predictive Analytics has great potential to increase the acquisition, development and retention of talent in an organisation.
On October 29 & 30, 2014, Python Predictions has developed the program for Predictive Analytics World London for the third time. We are very excited about this year’s line-up, with world-class analytics celebrities such as John Elder, Dean Abbott and David Hand, and company cases from Booking.com, Activision, Tata Sky, Verivox.de, NSA, the Belgian Government and DataKind UK.
Unigro (part of the Otto Group) is a Belgian distant-seller with a large product assortment. The company has a strategic focus on consumer (installment) credit, and we have been involved in predicting and monitoring their creditworthiness for more than 10 years. Since 2006, Unigro has realized a revenue growth of 25 percent whilst reducing credit risk, despite the financial and economic crises and the increase of riskier internet orders.
On October 6st 2014, Berten Van Herp joined our team as an Analyst. Berten holds a master’s degree in mathematics at the University of Brussels.
Berten started his professional career as Business Intelligence analyst, in charge for delivering custom-made analysis for pharmaceutical companies.
On September 25th, 2014, Python Predictions has organized its second closed customer workshop, with case studies from Eni, Optima and Unigro and participants from Eni (utilities), Optima (financial planning), Unigro (direct selling B2C), ING (banking), Manutan (direct selling B2C), bpost (postal services), AG Insurance (insurances) and Direct Social Communications (fundraising).
On September 1st 2014, Nele Verbiest joined our team as a Senior Analyst. Nele holds a master’s degree in mathematical computer science and a PhD in computer science, both from Ghent University.
In her PhD, Nele focused on data mining, and more specifically on nearest neighbor classification. Find Nele’s PhD here.
On June 4th 2014, Evi Van De Velde (Optima) and Pieter Van Bouwel (Python Predictions) have presented how Optima uses Predictive Analytics and IBM SPSS Modeler to efficiently attract new clients. Optima is the Belgian market leader in financial planning solutions. In collaboration with Python Predictions, Optima developed its first predictive models in support of its marketing decisions. This resulted in a tool that allows the company to single out prospects with the highest potential for its future campaigns. Find additional details about the event here.
On June 4th 2014, Geert Verstraeten has presented our journey with Unigro in credit scoring and fraud detection at SAS Analytics2014 in Frankfurt. Unigro (part of the Otto Group) is a Belgian distant-seller with a large product assortment. The company has a strategic focus on consumer (installment) credit, and has used SAS analytics technology for more than 10 years to predict and monitor creditworthiness. Using these techniques, since 2006, Unigro has realized a revenue growth of 25 percent whilst reducing credit risk, despite the financial and economic crises and the increase of riskier internet orders. Find additional details about the event here.
On March 18, 2014, Geert Verstraeten has presented a keynote speech at the Agoria ICT Trends Network Event, the networking event for top ICT management of the Belgian federation of the technological industry Agoria. In this speech, he offered his views on how to create business value from (big) data sources. Find the presentation here.
On January 13th 2014, Olivier Vermeersch joined our team as an Analyst. Olivier is a master in applied economics and an advanced master in marketing analysis at Ghent University. Before Python Predictions, Olivier was active as a customer intelligence analyst for Carrefour – a key player in the Belgian retail market.
On October 23rd & 24th, Geert Verstraeten organised and chaired Predictive Analytics World London for the second time. This edition featured analytics success stories (by Deutsche Bank, AOL, the Belgian government, BNP Paribas Fortis,…), surprising applications (e.g. stock market predictions, online & mobile campaigns) and managerial issues (e.g. creating a Predictive Analytics strategy, developing analytical talent). This edition also featured a keynote speech on Process Mining by domain expert Prof Van der Aalst.
In the fall of 2013, Wouter Buckinx has presented how eni Gas & Power uses analytical techniques to recover unpaid bills. By using predictive analytics, eni has made a remarkable improvement to its debt collection process. The story was presented on October 3rd, 2013 at the SAS Forum BeLux, together with Koen Ceulemans of eni, and on November 5th at Predictive Analytics World Berlin. Find the presentation here.
On 7 March 2013, Wouter Buckinx has presented a session on the MOA event entitled ‘Big data, great insights’ (in joint organisation with BAQMaR). In this introductory presentation, Wouter explained how Predictive Analytics can be used to retrieve insights from different (big) data sources, and how these learnings can be turned into value for customers and end-users.
In 2012, Geert Verstraeten was invited to co-organise and chair Predictive Analytics World in London. PAW is a unique event focussed on sharing the most successful applications of Predictive Analytics in business and government. Next to great case studies, the event featured specific methodological sessions as well as managerial sessions covering the challenges of attracting, managing and retaining analytical talent. In the event’s evaluation, over 95% of attendees claimed they would recommend the event to their peers.
In October 2012, Geert Verstraeten has presented our latest success story together with Ludo Longin, CEO at Direct Social Communications. This story was presented both at the SAS Belux Forum and at Dcongress, the largest yearly event of the Belgian Direct Marketing Association. In this presentation, we have demonstrated how Predictive Analytics was used to increase the efficiency of donor acquisition campaigns by 82%. Find a summary of our joint story here.
In May 2012, Geert Verstraeten was nominated for the first Direct WOW Award. With this award, the Belgian Direct Marketing Association aims to celebrate personalities with significant impact on the Direct Marketing domain in Belgium.
In different introductory sessions for IBM SPSS and for the Belgian BI-community, Wouter Buckinx has provided introductory workshops on how to use Predictive Analytics to boost marketing success in each phase of the customer lifecycle.
On a European roadshow (containing a.o. Predictive Analytics World in London, Analytics 2012 in Cologne and Analytics 2012 in Las Vegas, Geert Verstraeten and Wouter Buckinx have presented ING Belgium’s journey towards becoming a convinced user of Predictive Analytics, together with ING’s leading analysts Pieter Dyserinck and Jonathan Burez. In these presentations, we have covered ING’s journey from first successes to the integration of Predictive Analytics as the main engine of automated campaigns towards a large variety of target audiences and channels. For additional details, find our success story here.
On November 14th, Wouter Buckinx was featured Chairman of the Customer Intelligence Focus Day on the Loyalty World Conference in London. On this major event, Wouter led panel discussions with experts from the Royal bank of Scotland, Telefonica 02, Orange-France Telecom and Tesco.
In October 2011, Python Predictions became partner at the new Expertise Center for Database Marketing at IÉSEG School of Management, one of the top 10 business schools in France. IÉSEG ECDM is a research group that aims to establish durable research collaborations between academics active in research in the fields of database marketing and customer intelligence. The objective is to establish synergetic, innovative projects, offering novel solutions to real-life business needs related to database marketing.
In April 2011, Python Predictions celebrated its five years existence. In these five first years, we have executed projects for different key players in various industries, presented on several conferences in Belgium and abroad, and engaged in new analytical adventures in close cooperation with our clients, our partners and our research team. We have learned a great deal along the way, and are looking forward to the future.
On February 18th 2011, we have presented the results of our four-year collaboration with Overtoom Belgium, together with Overtoom’s Managing Director Ghislaine Duymelings and Marketing Director Jo Delange. In this presentation, we outlined how Predictive Analytics can serve as the driving force behind intelligent one-to-one communication – a process which leaded to an increase in marketing efficiency by 10%, and an increase in marketing relevance by 300%. The slides of this presentation may be found here
In September 2010, we were hired by Makro Cash & Carry Belgium with the goal of analyzing and segmenting their customer base according to customer profitability.
Makro is part of METRO Cash & Carry, an internationally leading self-service wholesaler. Makro services hoteliers, restaurants, caterers, canteens, small- and medium size retailers as well as service companies and businesses of all kind.
In August 2010, we have started our collaboration with Nuon Belgium (currently eni gas & power), an ambitious Dutch utility company formerly owned by the Swedish power company Vattenfall. Nuon produces, sells and delivers electricity, gas, heat and additional services to residential as well as professional customers.
In this first mission for Nuon, we succesfully increased Nuon’s prospection activities through the use of predictive analytics.
Following extensive training in SPSS Modeler, and after an interesting introduction with the SPSS community at SPSS Directions in Rome, we are proud to announce our first project using SPSS Modeler at Telenet.
Telenet is active in the Belgian market in internet and telephony services and comfort TV, offering solutions to residential as well as professional clients. In this project, we offer support in Telenet’s marketing intelligence department, focusing on campaign analysis and predictive model development.
On April 22nd, Geert Verstraeten has presented on the SAS-driven SAS-driven Spicing Up Analytics event in Tervuren. In this presentation, we have demonstrated a new Enterprise Miner node developed by Python Predictions – aimed at building models that combine predictive performance with intuitive business appeal.
Pieter Van Bouwel joined our team as Customer Intelligence Analyst. Pieter obtained an advanced Master in Marketing Analysis at Ghent University, and has previous working experience in telecommunications, utilities and retailing. Pieter will be active as a full-time analyst – with a focus on predictive modeling of customer behavior.
Python Predictions has received a ‘Rating 1-certificate’ from Dun & Bradstreet. This is the highest possible rating, accorded to about 15% of Belgian companies only. The D&B Rating is considered internationally as an objective indicator for the creditworthyness of companies. Only four years after our startup, we are truly proud of this proof of stability and trustworthyness of our company.
On October 22nd 2009, Python Predictions organised its first customer workshop. On this ‘closed-doors’ event, we disclosed detailed results of different implemented projects for Unigro, ING Belgium, La Poste, Mobistar and Overtoom Belgium. Topics covered succesful applications in the fields of predictive marketing modeling, credit scoring, customised offering and optimisation.
Sarah Timmermans joined our team as Customer Intelligence Analyst. Sarah obtained an advanced Master in Marketing Analysis at Ghent University, and has previous working experience in the financial sector. Sarah will be active as a full-time analyst – with a focus on predictive modeling of customer behavior.
On July 2nd, Geert Verstraeten has presented at A2009, the first European analytics conference organized by SAS. The conference brought together analysts and managers from all major industries including education, finance, government, healthcare, insurance, retail and pharmaceutical. In our presentation, we discussed the methods we use for combining predictive performance with interpretability, being a key success factor in our current projects, and we present an Enterprise Miner node we developed together with SAS Belgium for this purpose.
On May 7th 2009, Wouter Buckinx has presented on the 12th Relationship Marketing Conference of the Expert Group ‘Data Driven Marketing’ of the Belgian Marketing Foundation. In this presentation, we demonstrated how we tackle Overtoom Belgium’s ‘Customer-Product-Channel’ Challenge, and how predictive analytics were used to influence company results.
Florence Talbot joined our team as Customer Intelligence Analyst. Florence obtained an advanced Master in Marketing Analysis at Ghent University, and has previous working experience in the telecom sector. Florence will be active as a full-time analyst – with a focus on predictive modeling of customer behavior.
Kristof Coussement (PhD) joined our team as (part-time) Research Associate. Kristof is Assistant Professor of Marketing at the University College Brussels, where he teaches the principles of predictive modeling and market research. Kristof will be responsible for monitoring the latest scientific developments in the field of Customer Intelligence, fine-tuning existing methodologies and setting up innovative research projects.
We are proud to announce that, on 22 December 2008, we have initiated a project with De Post – La Poste, the Belgian organization responsible for the delivery of national and international mail, and one of the largest civilian employers in Belgium. In this project, we will be in charge of qualifying mail routes for the door-to-door activities, and we will perform a full assessment of the status and opportunities in terms of Customer Intelligence at De Poste – La Poste.
On November 5th 2008, we were invited to present at the SAS Forum Slovakia 2008. In this presentation, Wouter Buckinx explained how we increased the yearly revenue of Overtoom Belgium by over 10% on the long term, by using predictive marketing models.
On October 9th 2008, we have presented an insight into our work for Overtoom Belgium, a leader in the business-to-business retail market for office and warehouse supplies. In this presentation at the SAS BeLux Forum 2008, we have shown how we attained a proven yearly revenue increase of over 10%, by replacing the existing segmentation scheme by accurate predictive marketing models. We have illustrated the processes and methods used and the lessons learned during the project.
On March 11th 2008, we have discussed our views on the customer intelligence domain, together with Martine George of ING Belgium, at the Marketing Day 2008. In this presentation, it was our goal to inspire companies in different industries about the benefits of applying rigorous analytic techniques to enhance their Marketing ROI.
We are delighted that, in January 2008, Mobistar decided to hire our services in the field of customer intelligence analysis. The main focus for these analyses will lie on attrition and profitability issues of Mobistar’s B2B customers. One month later, we are excited with this new opportunity, and have currently achieved a fruitful start of the relationship. In the meanwhile, we are also proud that our relationships with ING Belgium, Overtoom and Unigro evolve with new, inspiring, challenges.
In November 2007, we have entered our ‘growth’ phase, by welcoming Toon Veldeman into our team. After graduating from the Master of Marketing Analysis at Ghent University in 2005, Toon has been working as a customer analyst in the financial services industry – both as an employee and as a consultant. Given his capabilities, motivation and personality, we are fully convinced that he provides a valuable enrichment to our team. We are very pleased that he decided to join us.
The presentation on October 11th 2007 was held by Wouter Buckinx and churn-expert dr Jonathan Burez of ING Belgium, and was entitled: “Avoiding the ‘Great Escape’ – Guidelines for developing actionable churn models”. The presenters offered support for their recommendations using various real-life business cases and (business and academic) literature.
On June 27th 2007, we have discussed the past, present and future of customer intelligence at ING Belgium into greater detail at the Stichting Marketing / Fondation Marketing Workshop on Marketing and CRM Analytics, together with Martine George of ING. In this presentation, it was our goal to inspire companies in different industries about the benefits of applying rigorous analytic techniques to enhance their Marketing ROI.
We are proud to announce that, in May 2007, we celebrated our one-year successful cooperation with ING Belgium. During this year, we had the opportunity to create predictive models of individual customer behavior for a large number of products and applications, in collaboration with the bank’s internal team of Customer Intelligence analysts. As a result of our successful cooperation with ING, we are proud to announce that SAS has recently published a customer success story covering our first accomplishments. Evidently, we are equally proud that our relationship with ING continues through new projects and challenges.
On Friday, February 2nd 2007, Python Predictions and Overtoom Belgium have signed a long-term agreement. Overtoom is part of the Manutan-group, one of the leaders in the business-to-business retail market for office and warehouse supplies. During the following nine months, our analysts will aim to improve the response rates on Overtoom’s direct mail campaigns towards customers and prospects. All analyses will be performed using SAS technology.
While companies are requiring the skills to predict what customers will spend in the future, only a few companies know how much their customers are spending with competitors. Our recently published study “Predicting customer loyalty using the internal transactional database”, Expert Systems with Applications, Volume 32, Issue 1, January 2007, Pages 125-134 focuses exactly on this topic.
Together with Senior CI Analyst Pieter Dyserinck of ING Belgium, Wouter Buckinx has presented ING’s most recent developments in Customer Intelligence on the SAS BeLux Forum on Thursday, October 12th, 2006 in Louvain-La-Neuve, Belgium. The launch of a Proactive Targeting Framework and the set up of a flexible CI toolkit enable ING Belgium to manage their current challenges. We show how the toolkit is designed, elaborate on the underlying technology, discuss which output measures have been observed for evaluation and demonstrate how the toolkit is used in practice in SAS Enterprise Miner.
As a logical evolution in our fruitful collaboration with SAS BeLux, we have become an official SAS Partner on August 1st, 2006. SAS is a leader in business intelligence software, having worldwide customers at 40,000 sites, including 96 of the top 100 FORTUNE Global 500 companies.
Together with dr. Bart Larivière of Ghent University, Geert Verstraeten has presented an overview of how SAS Software can be used for Customer Intelligence purposes in academic research. This presentation has taken place on Thursday, April 27th 2006 in Brussels, Belgium.
Geert Verstraeten has successfully defended his PhD in Applied Economic Sciences on Thursday, December 22nd 2005 at Ghent University, Belgium. His dissertation focuses on different issues in the predictive modeling of individual customer behavior, with applications in targeted marketing and consumer credit scoring.
Geert Verstraeten has presented his latest academic research on Saturday, November 19th at the 2005 International Workshop on Customer Relationship Management at New York University. His presentation covered the topic of predicting customer loyalty.
Wouter Buckinx and Geert Verstraeten have presented their research on predicting customer loyalty on Tuesday, October 25th 2005 in Essene, Belgium.
Wouter Buckinx and Geert Verstraeten have presented their latest academic research on Friday, July 15th 2005 at one of the largest and most respected conferences in Operations Research. Their presentations covered the topics of predictive modeling and targeting in the marketing domain.
Wouter Buckinx has successfully defended his PhD in Applied Economic Sciences on Monday, July 4th 2005 at Ghent University, Belgium. His dissertation is focused on optimizing the leaflet targeting policy of a major European retail chain. This system has proven to be highly successful, and is currently being implemented by the retailer.
Wouter Buckinx and Geert Verstraeten have been jointly selected as SAS 2005 (PhD) Student Ambassadors, and have presented their award-winning research project on Thursday, June 23rd 2005 in front of a business audience in Lisbon, Portugal. In this project, they illustrate that using the internal transactional database only, it is possible to determine how much your customers spend at competing stores. They also provide possible applications for the knowledge of customer loyalty.