How Tom won his first Kaggle competition

tom wins kaggle

This is a copy of Tom’s original post on Github.

Winning approach of the Facebook V Kaggle competition

The Facebook V: Predicting Check Ins data science competition where the goal was to predict which place a person would like to check in to has just ended. I participated with the goal of learning as much as possible and maybe aim for a top 10% since this was my first serious Kaggle competition attempt. I managed to exceed all expectations and finish 1st out of 1212 participants! In this post, I’ll explain my approach.


This blog post will cover all sections to go from the raw data to the winning submission. Here’s an overview of the different sections. If you want to skip ahead, just click the section title to go there.

The R source code is available on GitHub. This thread on the Kaggle forum discusses the solution on a higher level and is a good place to start if you participated in the challenge.


Competition banner

Competition banner

From the competition page: The goal of this competition is to predict which place a person would like to check in to. For the purposes of this competition, Facebook created an artificial world consisting of more than 100,000 places located in a 10 km by 10 km square. For a given set of coordinates, your task is to return a ranked list of the most likely places. Data was fabricated to resemble location signals coming from mobile devices, giving you a flavor of what it takes to work with real data complicated by inaccurate and noisy values. Inconsistent and erroneous location data can disrupt experience for services like Facebook Check In.

The training data consists of approximately 29 million observations where the location (x, y), accuracy, and timestamp is given along with the target variable, the check in location. The test data contains 8.6 million observations where the check in location should be predicted based on the location, accuracy and timestamp. The train and test data set are split based on time. There is no concept of a person in this dataset. All the observations are events, not people.

A ranked list of the top three most likely places is expected for all test records. The leaderboard score is calculated using the MAP@3 criterion. Consequently, ranking the actual place as the most likely candidate gets a score of 1, ranking the actual place as the second most likely gets a score of 1/2 and a third rank of the actual place results in a score of 1/3. If the actual place is not in the top three of ranked places, a score of 0 is awarded for that record. The total score is the mean of the observation scores.

Check Ins where each place has a different color

Check Ins where each place has a different color

Exploratory analysis

Location analysis of the train check ins revealed interesting patterns between the variation in x and y. There appears to be way more variation in x than in y. It was suggested that this could be related to the streets of the simulated world. The difference in variation between x and y is however different for all places and there is no obvious spatial (x-y) pattern in this relationship.

It was quickly established by the community that time is measured in minutes and could thus be converted to relative hours and days of the week. This means that the train data covers 546 days and the test data spans 153 days. All places seem to live in independent time zones with clear hourly and daily patterns. No spatial pattern was found with respect to the time patterns. There are however two clear dips in the number of check ins during the train period.

Accuracy was by far the hardest input to interpret. It was expected that it would be clearly correlated with the variation in x and y but the pattern is not as obvious. Halfway through the competition I cracked the code and the details will be discussed in the Feature engineering section.

I wrote an interactive Shiny application to research these interactions for a subset of the places. Feel free to explore the data yourself!

Problem definition

The main difficulty of this problem is the extended number of classes (places). With 8.6 million test records there are about a trillion (10^12) place-observation combinations. Luckily, most of the classes have a very low conditional probability given the data (x, y, time and accuracy). The major strategy on the forum to reduce the complexity consisted of calculating a classifier for many x-y rectangular grids. It makes much sense to make use of the spatial information since this shows the most obvious and strong pattern for the different places. This approach makes the complexity manageable but is likely to lose a significant amount of information since the data is so variable. I decided to model the problem with a single binary classification model in order to avoid to end up with many high variance models. The lack of any major spatial patterns in the exploratory analysis supports this approach.


Generating a single classifier for all place-observation combinations would be infeasible even with a powerful cluster. My approach consists of a stepwise strategy in which the conditional place probability is only modeled for a set of place candidates. A simplification of the overall strategy is shown below

High level strategy

High level strategy

The given raw train data is split in two chronological parts, with a similar ratio as the ratio between the train and test data. The summary period contains all given train observations of the first 408 days (minutes 0-587158). The second part of the given train data contains the next 138 days and will be referred to as the train/validation data from now on. The test data spans 153 days as mentioned before.

The summary period is used to generate train and validation features and the given train data is used to generate the same features for the test data.

The three raw data groups (train, validation and test) are first sampled down into batches that are as large as possible but can still be modeled with the available memory. I ended up using batches of approximately 30,000 observations on a 48GB workstation. The sampling process is fully random and results in train/validation batches that span the entire 138 days’ train range.

Next, a set of models is built to reduce the number of candidates to 20 using 15 XGBoost models in the second candidate selection step. The conditional probability P(place_match|features) is modeled for all ~30,000*100 place-observation combinations and the mean predicted probability of the 15 models is used to select the top 20 candidates for each observation. These models use features that combine place and observation measures of the summary period.

The same features are used to generate the first level learners. Each of the 100 first level learners are again XGBoost models that are built using ~30,000*20 feature-place_match pairs. The predicted probabilities P(place_match|features) are used as features of the second level learners along with 21 manually selected features. The candidates are ordered using the mean predicted probabilities of the 30 second level XGBoost learners.

All models are built using different train batches. Local validation is used to tune the model hyperparameters.

Candidate selection 1

The first candidate selection step reduces the number of potential classes from >100K to 100 by considering nearest neighbors of the observations. I considered the neighbor counts of the 2500 nearest neighbors where y variations are 2.5 times more important than x variations. Ties in the neighbor counts are resolved by the mean time difference since the observations. Resolving ties with the mean time difference is motivated by the shifts in popularity of the places.

The nearest neighbor counts are calculated efficiently by splitting up the data in overlapping rectangular grids. Grids are created as small as possible while still guaranteeing that the 2500 nearest neighbors fall within the grid in the worst case scenario. The R code is suboptimal through the use of several loops but the major bottleneck (ordering the distances) was reduced by a custom Rcpp package which resulted in an approximate 50% speed up. Improving the logic further was no major priority since the features were calculated on the background.

Feature engineering

Feature engineering strategy

Three weeks into the eight-week competition, I climbed to the top of the public leaderboard with about 50 features. Ever since I kept thinking of new features to capture the underlying patterns of the data. I also added features that are similar to the most important features in order to capture the subtler patterns. The final model contains 430 numeric features and this section is intended to discuss the majority of them.

There are two types of features. The first category relates to features that are calculated using only the summary data such as the number of historical check ins. The second and largest category combines summary data of the place candidates with the observation data. One such example is the historical density of a place candidate, one year prior to the observation.

All features are rescaled if needed in order to result in similar interpretations for the train and test features.


The major share of my 430 features is based on nearest neighbor related features. The neighbor counts for different Ks (1, 5, 10, 20, 50, 100, 250, 500, 1000 and 2500) and different x-y ratio constants (1, 2.5, 4, 5.5, 7, 12 and 30) resulted in 10*7 features. For example: if a test observation has 3 of its 5 nearest neighbors of class A and 2 of its 5 nearest neighbors as class B, the candidate A will contain the numeric value of 3 for the K=5 feature, the candidate B will contain the numeric value of 2 for the K=5 feature and all other 18 candidates will contain the value of 0 for that feature. The mean time difference between a candidate and all 70 combinations resulted in 70 additional features. 10 more features were added by considering the distance between the Kth features and the observations for a ratio constant of 2.5. These features are an indication of the spatial density. 40 more features were added in a later iteration around the most significant nearest neighbor features. K was set at (35, 75, 100, 175, 375) for x-y ratio constants (0.4, 0.8, 1.3, 2, 3.2, 4.5, 6 and 8). The distances of all 40 combinations to the most distant neighbor were also added as features. Distance features are divided by the number of summary observations in order to have similar interpretations for the train and test features.

I further added several features that consider the (smoothed) spatial grid densities. Other location features relate to the place summaries such as the median absolute deviations and standard deviations in x and y. The ratio between the median absolute deviations was added as well. Features were relaxed using additive (Laplace) smoothing with different relaxation constants whenever it made sense using the relaxation constants 20 and 300. Consequently, the relaxed mad for a place with 300 summary observation is equal to the mean of the place mad and the weighted place population mad for a relaxation constant of 300.


The second largest share of the features set belongs to time features. Here I converted all time period counts to period density counts in order to handle the two drops in the time frequency. Periods include 27 two-week periods prior to the end of the summary data and 27 1-week periods prior to the end of the summary data. I also included features that look at the two-week densities looking back between 75 and 1 weeks from the observations. These features resulted in missing values but XGBoost is able to handle them. Additional features were added for the clear yearly pattern of some places.

Weekly counts

Weekly counts

Hour, day and week features were calculated using the historical densities with and without cyclical smoothing and with or without relaxation. I suspected an interaction between the hour of the day and the day of the week and also added cyclical hour-day features. Features were added for daily 15-minute intervals as well. The cyclical smoothing is applied with Gaussian windows. The windows were chosen such that the smoothed hour, hour-week and 15-minute blocks capture different frequencies.

Other time features include extrapolated weekly densities using various time series models (arima, Holt-Winters and exponential smoothing). Further, the time since the end of the summary period was also added as well as the time between the end of the summary period and the last check in.


Understanding accuracy was the result of generating many plots. There is a significant but low correlation between accuracy and the variation in x and y but it is not until accuracy is binned in approximately equal sizes that the signal becomes visible. The signal is more accurate for accuracies in the 45-84 range (GPS data?).

Mean variation from the median in x versus 6 time and 32 accuracy groups

Mean variation from the median in x versus 6 time and 32 accuracy groups

The accuracy distribution seems to be a mixed distribution with three peaks which changes over time. It is likely to be related to three different mobile connection types (GPS, Wi-Fi or cellular). The places show different accuracy patterns and features were added to indicate the relative accuracy group densities. The middle accuracy group was set to the 45-84 range. I added relative place densities for 3 and 32 approximately equally sized accuracy bins. It was also discovered that the location is related to the three accuracy groups for many places. This pattern was captured by the addition of additional features for the different accuracy groups. A natural extension to the nearest neighbor calculation would incorporate the accuracy group but I did no longer have time to implement it.

The x-coordinates seem to be related to the accuracy group for places like 8170103882

The x-coordinates seem to be related to the accuracy group for places like 8170103882


Tens of z scores were added to indicate how similar a new observation is to the historical patterns in the place candidates. Robust Z-scores ((f-median(f))/mad(f) instead of (f-mean(f))/sd(f)) gave the best results.

Most important features

Nearest neighbors are the most important features for the studied models. The most significant nearest neighbor features appear around K=100 for distance constant ratios around 2.5. Hourly and daily densities were all found to be very important as well and the highest feature ranks are obtained after smoothing. Relative densities of the three accuracy groups also appear near the top of the most important features. An interesting feature that also appears at the top of the list relates to the daily density 52 weeks prior to the check in. There is a clear yearly pattern which is most obvious for places with the highest daily counts.

Clear yearly pattern for place 5872322184. The green line goes back 52 weeks since the highest daily count

Clear yearly pattern for place 5872322184. The green line goes back 52 weeks since the highest daily count

The feature files are about 800MB for each batch and I saved all the features to an external HD.

Candidate selection 2

The features from the previous section are used to generate binary classification models on 15 different train batches using XGBoost models. With 100 candidates for each observation, this is a slow process and it made sense to me to narrow down the number of candidates to 20 at this stage. I did not perform any downsampling in my final approach since all zeros (not a match between the candidate and true match) contain valuable information. XGBoost is able to handle unbalanced data quite well in my experience. I did however consider to omit observations that didn’t contain the true class in the top 100 but this resulted in slightly worse validation scores. The reasoning is the same as above: those values contain valuable information! The 15 candidate selection models are built with the top 142 features. The feature importance order is obtained by considering the XGBoost feature importance ranks of 20 models trained on different batches. Hyperparameters were selected using the local validation batches. The 15 second candidate selection models all generate a predicted probability of P(place_match|data), I average those to select the top 20 candidates in the second candidate selection step.

At this point I also dropped observations that belong to places that only have observations in the train/validation period. This filtering was also applied to the test set.

First level learners

The first level learners are very similar to the second candidate selection models other than the fact that they were fit on one fifth of the data for 75 of the 100 models. The other 25 models were fit on 100 candidates for each observation. The 100 base XGBoost learners were fit on different random parts of the training period. Deep trees gave me the best results here (depth 11) and the eta constant was set to (11 or 12)/500 for 500 rounds. Column sampling also helped (0.6) and subsampling the observations (0.5) did not hurt but of course resulted in a fitting speed increase. I included either all 430 features or a uniform random pick of the ordered features by importance in a desirable feature count range (100-285 and 180-240). The first level learner framework was created to handle multiple first level learner types other than XGBoost. I experimented with the nnet and H2O neural network implementations but those were either too slow in transferring the data (H2O) or too biased (nnet). The way XGBoost handles missing values is another great advantage over the mentioned neural network implementations. Also, the XGBoost models are quite variable since they are trained on different random train batches with differing hyperparameters (eta constant, number of features and the number of considered candidates (either 20 or 100)).

Second level learners

The 30 second level learners combine the predictions of the 100 first level models along with 21 manually selected features for all top 20 candidates. The 21 additional features are high level features such as the x, y and accuracy values as well as the time since the end of the summary period. The added value of the 21 features is very low but constant on the validation set and the public leaderboard (~0.02%). The best local validation score was obtained by considering moderate tree depths (depth 7) and the eta constant was set to 8/200 for 200 rounds. Column sampling also helped (0.6) and subsampling the observations (0.5) did not hurt but again resulted in a fitting speed increase. The candidates are ordered using the mean predicted probabilities of the 30 second level XGBoost learners.

Analysis of the local MAP@3 indicated better results for accuracies in the 45-84 range. The difference between local and test validation scores is in large part related to this observation. There seems to be a trend towards the use of devices that show less variation .

Local MAP@3 versus accuracy groups

Local MAP@3 versus accuracy groups


The private leaderboard standing below, used to rank the teams, shows the top 30 teams. It was a very close competition in the end and Markus would have been a well-deserved winner as well. We were very close to each other ever since the third week of the eight-week contest and pushed each other forward. The fact that the test data contains 8.6 million records and that it was split randomly for the private and public leaderboard resulted in a very confident estimate of the private standing given the public leaderboard. I was most impressed by the approaches of Markus and Jack (Japan) who finished in third position. You can read more about their approaches on the forum. Many others also contributed valuable insights.

Private leaderboard score (MAP@3) - two teams stand out from the pack

Private leaderboard score (MAP@3) – two teams stand out from the pack

I started the competition using a modest 8GB laptop but decided to purchase a €1500 workstation two weeks into the competition to speed up the modeling. Starting with limited resources ended up to be an advantage since it forced me to think of ways to optimize the feature generation logic. My big friend in this competition was the data.table package.

Running all steps on my 48GB workstation would take about a month. That seems like a ridiculously long time but it is explained by the extended computation time of the nearest neighbor features. While calculating the NN features I was continuously working on other parts of the workflow so speeding the NN logic up would not have resulted in a better final score.

Generating a ~.62 score could however be achieved in about two weeks by focusing on the most relevant NN features. I would suggest to consider 3 of the 7 distance constants (1, 2.5 and 4) and omit the mid KNN features. Cutting the first level models from 100 to 10 and the second level models from 30 to 5 would also not result in a strong performance decrease (estimated decrease of 0.1%) and cut the computation time to less than a week. You could of course run the logic on multiple instances and further speed things up.

I really enjoyed working on this competition even though it was already one of the busiest periods of my life. The competition was launched while I was in the middle of writing my Master’s Thesis in statistics in combination with a full time job. The data shows many interesting noisy and time dependent patterns which motivated me to play with the data before and after work. It was definitely worth every second of my time! I was inspired by the work of other Kaggle winners and successfully implemented my first two level model. Winning the competition is a nice extra but it’s even better to have learnt a lot from the other competitors, thank you all!

I look forward to your comments and suggestions, please go to my original post to post your comments.


Data Innovation Summit 2016 – Uitnodiging


Tweede ‘Data Innovation Summit’ brengt alle Belgische actoren uit data innovatiesector samen met onder meer Saskia Van Uffelen (CEO Ericsson), Sonja Rottiers (CFO AXA) and Annemie Depuydt (CIO KULeuven) , 7 tracks en meer dan 40 presentaties.

Afspraak dinsdag 10 mei 2016 vanaf 8 uur bij AXA, Vorstlaan 25, Watermaal-Bosvoorde – Brussel

Na de succesvolle eerste editie van vorig jaar die meer dan 500 toeschouwers lokte, sprak het voor zich dat organisator Philippe Van Impe hier een vervolg aan zou breien. Hij kopieert het concept van vorig jaar echter niet klakkeloos, maar gooit de verhouding mannelijke en vrouwelijke sprekers radicaal om. Hiermee wil Van Impe aantonen dat ook vrouwen sterk betrokken zijn in deze wereld die op het eerste zicht hoofdzakelijk mannelijk aandoet.

Deze tweede Data Innovation Summit brengt alle actoren uit de Belgische data innovatiesector samen. Meer dan veertig sprekers uit de politieke, academische en bedrijfswereld presenteren hun bijdrage om België in de data innovatiemarkt te positioneren. Ook startups krijgen er een platform om hun big data innovaties te showcasen.

De Data Innovation Summit is een breed gedragen initiatief dat georganiseerd wordt door de European Data Innovation Hub (+1500 leden / Brussels Data Science Meetup / @datasciencebe).


Met deze conferentie wil de European Data Innovation Hub een stand van zaken bieden over Belgische data innovatie en alle actoren oproepen om de krachten te bundelen en big data, open data en data innovatie hoger op de zakelijke en politieke agenda te plaatsen.


Praktische informatie

Meest recente update van de agenda en informatie over sprekers vind je op

Alle presentaties zullen na de conferentie beschikbaar zijn.

Volg de Data Innovation Summit ook via Twitter #DIS2016


Contact & organisatie

European Data Innovation Hub

Philippe Van Impe
0477 23 78 42

DiS16_Poster_v02 (1) (1)

Event – DIS2016 – Data Innovation Summit – March 23rd – AXA Brussels

#DIS2016 – AXA building – Brussels – March 23rd.
Reserve your seat:

Datascience innovation summit

The main theme will be: “Digital Transformation”

Last year’s success motivated us to hold The Data Innovation Summit again in the AXA building in Brussels.

We expect over 500 visitors interested in #DataInnovation, #DataScience, #BigData coming from the public, academic, business and startup world.

We will have less presentations than last year and provide you with more time to network and have one-to-one sessions.


Reserve your seat now for DIS2016 – 23rd March 2016 Brussels.

Registration fee starts at just over 25€ if you are an early bird, you finalise the survey and share your enthusiasm with on your linkedin account…



parallel tracks

The list of exhibitors, partners and sponsors is available here.

#DIS2016 – Survey


  • You will get 100€ discount when :

– you respond to the survey.
– share your enthusiasm in a Linkedin post

Possible discounts:

Contact to request your discount

  1. Unemployed professionals and carreer switchers get 75% discount
  2. Active members of the community who have done presentations at other meetups or shared their knowledge in trainings and coaching sessions get 50%
  3. Co-workers and startups hosted at the HUB get also 50% discount

Partnering with this event:

We welcome partners, you can review the sponsoring options here.

Remember DIS2015:

  • Here are some of the videos of last year’s event.
  • Some pictures of last year’s event



List of Open Data Portals from Around the World by


We will have a special meetup about Using Open data to promote Data Innovation on Thursday, November 26, 2015 @VUB. Here is a Comprehensive List of Open Data Portals from Around the World. is the most comprehensive list of open data portals in the world. It is curated by a group of leading open data experts from around the world – including representatives from local, regional and national governments, international organisations such as the World Bank, and numerous NGOs.

The alpha version of was launched at OKCon 2011 in Berlin. We have plenty of useful improvements and features in the pipeline, which will be launched over the coming months.

If you have an idea for a feature you think we should add, please let us know via the discussion forum.

open data

The Open Definition sets out principles that define “openness” in relation to data and content.

It makes precise the meaning of “open” in the terms “open data” and “open content” and thereby ensures quality and encourages compatibility between different pools of open material.

It can be summed up in the statement that:

“Open means anyone can freely access, use, modify, and share for any purpose(subject, at most, to requirements that preserve provenance and openness).”

Put most succinctly:

“Open data and content can be freely used, modified, and shared by anyone forany purpose

Event – Official Opening of the European Data Innovation Hub in Brussels – October 20th 10AM

HUBdatainnovation invitation

We are so pleased to announce that our Hub will officially be inaugurated by Alexander De Croo, Bianca Debaets, Frank Koster, Marietje Schaak and Jörgen Gren. Over 100 directors and managers have already confirmed their presence to this ribbon cutting ceremony that will be held in our new offices in the Axa building.

266px-Alexander_de_croo_675 bianca-in-brussel frank marietjejorgen_gren

Join us on October 20th at 10AM to meet young data startups, talk to representatives of the academic world, share your ideas with the politics representatives that support us and discover our datascience training offering.

There are only a few more spaces left so please hurry and reserve your seat via our eventbrite page.

I’m looking forward to meeting you in our new offices soon,

Philippe Van Impe
asbl European Data Innovation Hub vzw 
Inspire – Innovate – Connect

Call for papers – #BDW – Using Open data to promote Data Innovation


Our November 26 meetup will be an event part of the international Big Data Week  #BDW15 , a global festival of data.

We will focus on Open Data:

 ‘Open data as enabler for improved decision making and new product and service offerings.’

Please contact if you would like to share your experience and submit your subject.

The call for papers is now open.

Potential Agenda:

  • How government can promote open data. 
    Open data has the potential not only to transform every sector of the economy but also to unleash billions in global economic value annually. Government has a critical role to play.
  • The Belgian open data initiative.
  • EC Open Data Policy
  • Open data as enabler for improved decision making and new product and service offerings.

Using Open data to promote Data Innovation

Thursday, Nov 26, 2015, 6:30 PM

VUB – Aula QB
Pleinlaan 2B – 1050 Brussels, BE

23 Business & Data Science pro’s Attending

This event is part of the #BDW15 , a global festival of data.Our topic will be Datasciences and Open Data: ‘Open data as enabler for improved decision making and new product and service offerings.’The call for papers is now open, please contact [masked] to submit your subject.Agenda:• How government can promote open data.  Open data h…

Check out this Meetup →

#datascience training programmes

Why don’t you join one of our  #datascience trainings in order to sharpen your skills.

Special rates apply if you are a job seeker.

Here are some training highlights for the coming months:

Check out the full agenda here.

Leo De Bock brings the speech of Minister Kris Peeters during the Data Innovation Summit

Leo De Bock Leo De BockLeo De Bock 1

It is so nice to be supported by Minister Kris Peeters and his team during our first Data Innovation Summit. Thank you Leo for the excellent presentation.

  • Here is the video from Leo De Boeck.
  • See all presentations from the summit here
  • See all  the pictures from the event here


Ladies and gentlemen, 

First of all, I would like to apologise Vice Prime minister Kris Peeters for not being able to make it to this meeting due to a political meeting that succeeded to eventually dominate his agenda.  

On days like these, when the bright minds of the world of data come together, we have a great opportunity to look ahead. We can discuss the next step in a field that progresses at an impressive speed. 

Now that digital has become the new normal, what will be the new extraordinary?  

The digital economy is one of the most dynamic and promising sectors in terms of development potential. Its possibilities for growth far exceed those of other sectors. Mobile data traffic doubles each year, the use of the internet does so every two or three years. Today, 4 million people work in ICT in the European Union and their share increases with 3% per year, despite the economic crisis. 

For the federal government of Belgium, our prime goal is to translate this digital growth into job creation. This is why we are developing our digital agenda. If we want to stay ahead as a digital nation, we’ll need to invest. We have a more developed high speed internet infrastructure than most other countries. We do not want to give up on that advantageous position.  We instead wish to continue to invest in a 4G and LTE advanced network in Belgium. 

At the same time, we need to invest in our regulatory framework. We need to update our privacy legislation that dates back to 1992, which, in digital terms, is the stone age. Privacy is a key driver for digital progress. The digital revolution will come to a halt when people’s trust diminishes. People have to be granted the right to information ànd the right to be overlooked, ignored, forgotten. Moreover, people need to feel safe when they go online. 

The government therefore, wants to make the CyberSecurityCenter operational this year. This center will work out a strategy to secure our nation’s digital network and the information we have online. Moreover, the Federal Public Service for Economy organizes campaigns to raise awareness on the data and digital fingerprint that people leave online.  

In today’s age of big data, these kinds of campaigns have become a necessity. Never before in the history of mankind, data have been collected, processed and linked at this massive a scale. A company’s power today is not just valued in terms of capital, but also in terms of data. Data is the new gold. Data means a company can produce, ship and market their products and services far more efficiently. 

In fact, our cabinet has already reaped the benefits of big data. We have been working together with a promising Belgian start-up, called Dataminded, who analyse our social media activities and we have already changed our communication policy on the basis of this analysis. 

Both the Vice Prime Minister and his staff are convinced of the great added value of big data. But, ladies and gentlemen, this does not mean that all bets are off. As I mentioned earlier, we need some form of regulation. Moreover, we need to think about creating a level playing field. 

The term ‘big data’ is quite self-explanatory. Big companies, first and foremost, are capable of gathering a critical mass of data for useful analysis. Those companies have the means to buy data or to invest in data processing. The question therefore is, how we can make sure that small and medium sized companies – those that somewhat are the backbone of our economy – can also become part of the big data story.  

This is why the federal government is working on legislation that makes open data accessible to citizens, companies and researchers. The exchange of data between governments and other organisations, will strengthen literally every citizen and company. Sharing data means strengthening everybody. 

Companies that own data, need to keep this in mind. Big data should be more than a new way to maximize profits. Big data should also benefit society. Take product information for instance. It is abundantly present in digital networks and it is used to reduce costs, boost productivity and make marketing efforts more efficient. Product information is often highly specialized, technical and exhaustive. It is so exhaustive that it confuses the average consumer. This is the point where big data should cross the boundary between economic logic and social logic. 

Let me give you one very concrete example. In December of last year, a new European regulation on Food Information to Consumers came into force. This legislation will ensure that consumers get more information on the food products that are put on sale in stores. But given the abundance of information, it is difficult for the consumer to use this in a meaningful way. So wouldn’t companies rather share their product information with others who can present these data in a more comprehensive way? Isn’t is socially responsible, isn’t it a corporate social duty for them to share information so that this legislation can actually be applied and we can make the consumer more aware, give him a chance to make rational decisions? This is where data should be turned into knowledge. 

Now, some companies will consider this a threat. But frankly, they are wrong. If you take initiative, you create opportunities. You get ahead instead of trailing the pack. First of all, transparency about data creates trust. And trust boosts business. Secondly, when companies provide their data directly, they can be sure that the data on the market are correct. Companies that continue to shield and hesitate and stay aloof, make the wrong choice. Because eventually the data will see the light of day. The huge multinational internet companies will put this data on the market sooner or later. And they will not wait for an agreement or cooperation. 


Ladies and gentlemen, 

Big data is indeed big business. But it also means big responsibility. While having the new gold in your hands, you should think twice about what you use it for. Eleanor Roosevelt once said that “if you need to handle yourself, you should use your head; but if you have to handle others, use your heart.” 

I hope hearts and minds will work together when we develop the new extraordinary. Because that is what big data is. Now that digital has become the new normal, big data is the next leap.  

For now, I wish all of you a productive and fruitful conference. And together with Vice Prime Minister Peeters, I do look forward to the great innovations that all of you, the bright minds of big data, will create in the years to come. 

I thank you.


Thank you for making the Data Innovation Summit a success

sponsorsThank you all for your engagement and active participation to our Data Innovation Summit.

Yesterday 68 presentations were delivered on time allowing over 400 data lovers to have enough time to network and share ideas with their peers.

This would not have been possible without a professional team of volunteers, a team of friends making the craziest schedules possible.

I would like to thank AXA and our sponsors for supporting us.

The speakers were amazing, tortured to accept the most horrible presentation format called ignite and delivering it with so much grace and passion, beautiful.

What a pity that we could not schedule all presentations and that we had to turn down so many participants because the event was sold out. We will handle this differently next time.

Together we have reached our first milestone yesterday, it is time to wake up now and work together to accelerate the development of new projects that will position us better in this merciless digitalization race. Let’s bundle our energy and put Belgium back on the map of most innovative countries, the place where it is so easy to start-up a company.

Your turn now, please give us your feedback about our Summit:

The pictures taken at the summit will be available on

The presentations video’s will also soon be available on

Have you seen the analysis of the results (over 600) of the survey made by Ward, Dieter & Nicholas, Nele and Rik. I’m looking forward to these presentations during the finals of April 16th.

Thanks again,

Philippe Van Impe

Ready for the Data Innovation Summit, Survey & Challenge ?


Your community is holding its first full day summit this week on Thursday in Brussels.

We have received so much support from all parts of Belgium. I’m so proud of all our actors contributing to this amazing agenda. It will be a fast paced day with over 50 presentations and over 30 exhibitors. You will see and meet:

  • Most important Universities of Belgium
  • Major corporations will explain how they prepare to become data driven
  • Political powers represented by Kris Peters
  • Over 20 start-ups
  • Over 480 data experts
  • Over 30 exhibitors

Here are some Highlights of the day are:


  • The presentation of Kris Peeters & Vincent Blondel
  • The presentations of Reservoir Lab from UGent. 100 k$ winner of the international Kaggle competition.
  • The launch of the European Data Innovation Hub.

The challenge

If you have already registered and are on the waiting list, take the survey that will lead you to your ticket directly. Once you have answered this short poll we would like to challenge you and your team to analyze this data and to prepare a presentation for our finals.

About the summit:

  • We open our doors at 07:30. The Axa building is situated Boulevard du Souverain, 25. 1170 Bruxelles.
  • A parking space has been reserved for you. Use the Tenreukenlaan to proceed to the visitors parking of Axa.
  • The venue is 10 minutes walk from metro Hermann Debroux.
  • All presentations will be recorded and made available on our website.
  • Here is the list of participants .
  • Press Coverage: We have issued a press release

I’m looking forward to meeting you on Thursday.

Philippe Van Impe

%d bloggers like this: