Improving your Consulting Skills as Data Scientist

While developing technical skills is vital for Data Scientists, ensuring to clearly communicate procedures and findings is equally important. For this purpose, we invite you to join our upcoming training “Consulting Skills for Data Scientists”, delivered by Eric Lecoutre and Sven Wauters, from WeLoveDataScience.

You can find more details in the related event. There is a special offer for students, jobseekers, and our partner companies. Don’t hesitate to join this training!

Data Analysis Internship at a political party

One of the participants at the Summer Coding Camp sent us this opening from his organisation:

The Brussels-based ALDE Party is seeking for a Data Analysis trainee for a 6 month period. The candidate should have a background in Statistics or Computer Science, or from the fields of Economics, Political Science, Psychology, or Sociology, with strong quantitative abilities.

More details can be found here: https://www.aldeparty.eu/data-analysis-trainee-mf

Please feel free to share this opportunity!

Kris Peeters @ DISUMMIT 2017: “Success and failure in big data projects”

DiS17_Speaker_11:30_Kris Peeters

Kris Peeters is the founder and CEO of Data Minded. He leads a group of passionate tech experts working in the area of Data Engineering. Working on the implementation of Big Data projects has taught him some best practices, which he is always glad to share.

About Kris and his tips to avoid Murphy’s law in big data projects

Kris started his presentation with a case in which the effort of several months was simply replaced by 3 Excel sheets. This was the perfect way to exemplify the frustration that all people working in big data projects may experience, pointing out the need of listening to the advice of those who have gone through the process.

With that kick-off, Kris shared which steps would ensure that Murphy law could step into the process. He started pointing the fact of performing long development cycles without performing some testing nor sharing it with the customer, and also highlighted how having a “golden hammer” (well known tool) giving the wrong image of taking success for granted. Dismissing the importance of diversity (in terms of including different backgrounds) and proper coaching are also red flags to keep an eye on.

Besides his honest, incisive feedback, what is truly appreciated from Kris is his interest to see things from different angles, and to be very inclusive on the perspective of other stakeholders. At the end, knowing how to use big data cannot go too far without the qualitative input that the customer can offer.

So… before kicking off a project, it is crucial to spot the involved weaknesses, as this allows  assessing them and being proactive about them.

A takeaway from Kris’s presentation

“When kicking off a project, do a pre-mortem to spot which risks may appear”

Thank you, Kris, for the useful tips! We are sure that they save some difficult moments to many people.

Interview:

Presentation:

 Deck:

[To be uploaded asap]

Presentation’s sketch:

DIS17_KrisPeeters

Sandro Sinigaglia @ DISUMMIT 2017: “When data and analytics exposes the bad guys?”

DiS17_Speaker_Sandro Sinigaglia

Sandro Sinigaglia is the head of the e-Fraud Competence Center at ING Belgium. He also is a topic leader and lecturer at Solvay Brussels School (ULB) and at international cybersecurity conferences.

Danny Moerenhout is an Expert Process Manager within the Competence Center e-Fraud at ING Belgium.

About Sandro and Danny’s “cat and the mouse” game with cyber criminals

When Sandro and Danny started talking about cybersecurity, the stage became a play simulating the real situations banks are now facing when cybercriminals and black hat hackers pop into scene.

The shift of online banking has opened the doors to create ways to deceive users. Actions like the creation of mule accounts and the proliferation of “spray and pray” malware initially led the banks to work on a sniper view basis, spotting and performing an action to stop a specific criminal.

However, as criminals have been becoming more savvy on banks practices and have reacted accordingly, the bank’s strategy has evolved from sniper to a 360° view,  in which data is the key element to spot patterns and continuously improve list management and rules.

Besides offering a humorous view on the “cat and mouse” game that banks and criminals play, Sandro and Danny made clear that organisations need to abandon a classical segmentation approach and focus more on building algorithms based on behaviour, allowing to connect the dots with social media.

Our favorite phrase from Sandro and Danny’s presentation

“On the internet, nothing is what it seems”

 

We look forward to meet Sandro and Danny in the future so they can share more insights on the fight against cybercriminals!

Sandro and Danny’s interview:

[to be added asap]

Sandro and Danny’s presentation recording:

Sandro and Danny’s deck:

[to be added asap]

Sandro and Danny’s presentation drawing:

DIS17_SandroSinigagua

Serge Masyn @ DISUMMIT 2017: “Boost Innovation: Convince your CEO to run a Hackathon” ”

DiS17_Speaker_09:42_Serge Masyn

Serge Masyn is a Director at Johnson and Johnson, where he has been working for almost 20 years, being his current focus the design and implementation of positive identification methodologies of vaccines. Serge believes in collaboration with other organisations as a key factor to achieve improvements in public health.

About Serge  and his experience organising the Dengue Hack

Serge explained to us how viral diseases like dengue pose higher risks than in the past. The evolution in factors like climate change, migration and movement of goods has contributed to the spread of viral diseases like dengue.  Due to the global warming, there are more warm zones where the mosquito can lay its eggs. Derived from mobility, when there was a dengue outbreak in Madeira, the disease soon reached Portugal, as there is constant number of people traveling between the 2 countries. The transportation between continents of goods like tyres (which may contain mosquito’s eggs) increases the possibility that dengue can appear in new regions.

Due to these circumstances, the WHO forecasts that a 50% of the world population is subject to contract it. Hence, it is crucial to monitor the channels in which it is spread and to forecast the outbreaks location and magnitude. This can only be achieved with an analysis of broad and detailed sets of epidemiological, weather, movement of people, movement of tyres, construction sites data. And that is when the idea of a hackathon popped into scene, and Serge got in touch with Philippe, from the Data Innovation Hub, to organise it.

While Serge’s main objective was to crowdsource a solution, he soon realise that the benefits of a hackathon were beyond that. Besides putting together creative minds pooling different approaches to predict dengue outbreaks, it also allowed expanding the company’s network, working with innovative software, and raising awareness on both public health and on the importance of dengue.

Serge wrapped up our interview with the reflection that organising a hackathon brings diversity in perspectives… and that yes, he is open to repeat the experience!.

Our favorite phrase from Serge’s presentation

“Think out of the box, by thinking out of the box”

 

We look forward to see that, same as Serge, many other companies realise the benefit of organising hackathons!

Serge’s interview:

Serge’s presentation recording:

Serge’s deck:

[To be uploaded asap]

Serge’s presentation drawing:

DIS17_SergeMasyn

Geert Verstraeten @ DISUMMIT 2017: “Predicting Absenteeism”

DiS17_Speaker_Geert Verstraeten

Geert Verstraeten is Managing Partner at Python Predictions, a Belgian niche player with expertise in the domain of Predictive Analytics. His ambition is to coach analysts and their managers to become successful with predictive analytics.

About Geert view on how absenteeism can help predicting burn-out

In the current job scenario, in which 2 of 3 workers experience stress, it is crucial to predict when can this situation escalate to burn-out, generating a high financial cost for the companies and having a negative impact on the employee’s health and motivation.  This motivated Geert to develop a model to predict burn-out basing on absenteeism.

Framing the problem in the proper way is key to use data  effectively. In this case, it involved focusing on absenteeism using payroll data and linking it to evaluation scores. This model works best when aggregated to teams, allowing organisations to have a better panorama on how their organisation is performing to build a strategy integrating different departments.

In an era in which the word “burn-out” is frequently mentioned (and feared), it is encouraging to see how data coming from different sources can be used to build healthier job environments.

Our favorite phrase from Geert’s presentation

“Make sure that you are solving  the right problem”

 

We hope that Geert’s insights can be used in a broader way for spotting burn-out!

Geert’s interview:

Geert’s presentation recording:

Geert’s deck:

[To be uploaded asap]

Geert’s presentation sketch:

DIS17_GeertVerstraeten

Pierre-Nicolas Schwab @ DISUMMIT 2017: “What we desperately need are ethical algorithms”

DiS17_Speaker_11:15_Pierre-Nicolas Schabb

Pierre-Nicolas Schwab is in charge of the Big Data/CRM program at RTBF, the French-speaking public broadcasting organization of Belgium. He holds an MSc in chemistry, a MBA in finance and a PhD in marketing.  As external consultant, he has carried out some 70 assignments in various industries, combining his passion for data and deep interest for consumer behaviors.

About Pierre and setting up rules to have more ethical algorithms

Pierre’s passion for linking data with consumer behaviour led him to share with us why there should be more focus to make algorithms more ethical. While algorithms are present everywhere, people have are just questioning their integrity. Little by little, the dark side of algorithms is becoming more visible, and -most likely- in a deliberate way.

The fact that organisations are using algorithms against the benefit of their user base raises eyebrows and threatens its relationship. For instance, in an era in which “fake news” is a common topic, personalising the content may be perceived as hiding some key facts which provide a full view of the context, breaking the users’ trust. This is crucial for companies like RTBF.

The reality is that algorithms are not neutral, and either by their technical limitations or by the vision of their developers, they pose serious threats. At the end, it is essential to set up rules towards algorithms design and use. This means that the whole algorithm cycle (design, implementation and follow-up) is subject to a transparency and governance process supported by showing the algorithms (taking them out from their “black box”) and educating the users on how the algorithms work.

Organisations need to include rules and procedures towards algorithms. This is key to ensure both a healthy reputation and relationship with their users.

Our favorite phrase from Pierre-Nicolas’s presentation

“Cow-boy behaviors must end!”

 

We thank Pierre for the handy tips he shared to help organisations to aim for more ethical algorithms, hoping that many organisations consider them!

Pierre’s interview:

Pierre’s presentation recording:

Pierre’s deck:

[To be uploaded asap]

Pierre’s presentation drawing:

DIS17_Pierre-NicolasSchwab

Jos Polfliet @ DISUMMIT 2017: “Preventing suicide using text analytics”

Jos Polfliet is a data scientist working for SAS who loves analysing complex problems and predicting the future. He believes in using the power of data to transform and improve the world in almost every field, as he has experienced with the project he came to share at DI Summit.

About Jos sharing how text analytics can prevent suicide

Jos shared his experience in Canada at that the Data Impact Challenge organised by the Canadian organisation Imagine Nation. The challenge Jos and his worked on was on  analysing 1.1 million tweets of Canadian teenagers on bullying and suicidal thoughts to spot patterns which could then be used to develop actions towards youth mental health.

Rather than the award his solution received, what is focal to Jos is how can it be replicated in other locations. During his presentation at DI Summit, Jos requested the attendees to provide leads of people who could help him to replicate the analysis in Belgium. Less than an hour later, he already had 3 leads, some of them related to profiles Jos would not have thought about!

In order to boost the power of data, Jos mainly gets inspired by following people online. He attends offline events mainly for meeting potential partners for projects, and -like he did- to make a call for a specific need.

To wrap up, Jos invited organisations to reflect on the data they have and organise regular review with the business departments to, according to their priorities, crosscheck with the business departments to spot the opportunities and projects which can be developed.

Our favorite phrase from Jos’s presentation

“Don’t let data sitting around without asking what can be done with it”

 

We hope that Jos keeps on sharing with us how data analysis has been used in different contexts!

Jos’s interview:

Jos’s presentation recording:

Jos’s decks:

[To be uploaded asap]

Jos’s presentation drawing:

DIS17_JosPolfliet_sketch

Rishi Kumar @ DISUMMIT 2017: “Data for Good, Lesson from the Frontline”

DiS17_Speaker_09:28_Rishi KumarRishi Nalin Kumar is the Chief Scientist at eBench. He is an analytics expert who has worked in the corporate and startup environments while also supports DataKind, which fosters the collaboration between top data scientists and social change organisations in projects to maximise positive social impact.

About Rishi and what ideal data for good projects should have

Rishi visited us from London to share his main lessons as DataKind’s Chapter Lead. He is enthusiastic about having an increasing number of initiatives (like hackathons and organisations -of all kinds-) which have as startpoint a societal challenge requiring solutions via analysing data pools.

His experience in building synergies to use data as  a catalyser for social change has taught him to keep an open mind to changes, as the original objective set for a project may evolve along with its findings. Additionally, Rishi emphasized the importance of involving people with a non-data background in the projects in order to build well-grounded solutions. What is also important to consider is that a project’s good cause requires many other element on board to be successful, and that it is essential to always validate that the following elements are on board:

Relevant data + reliable partner to work with

+ well framed problem + simple solution to be implemented

When asked about what is a major general challenge towards data, Rishi confirmed that  organisations need to update their view towards customer segmentation. The usual approach of defining a very fixed set up rules (for instance, grouping customers by the last way they performed a certain action) is less efficient than considering identity traits to spot behavioural patterns. The complexity involved to do so would certainly worth approaching customers in a more assertive manner.

A takeaway from Rishi’s presentation

“If data is the new oil, privacy is the new climate change”

 

Looking forward to keep on listening to your lessons on using data for good, Rishi!

Rishi’s interview:

Rishi’s presentation recording:

Rishi’s deck:

Rishi’s presentation’s sketch:

DIS17_RishiNalinKumar

Filip Maertens @ DISUMMIT 2017: “Can A.I. help us build a better world?”

DiS17_Speaker_09:14_Filip Maertens

Filip Maertens is a cyber-security expert. He has worked in the military, intelligence, critical infrastructure and banking industry; both in engineering and project management roles. Besides loving to code software, he also loves dogs, and climbing mountains. You can find more over Filip by visiting his blog: https://filipmaertens.com

About Filip’s view on the challenges Artificial Intelligence is facing

Filip kicked off his presentation highlighting how society (ourselves!) have generated a science-fiction fatalist scenario on Artificial Intelligence’s effects, without realising that we are already in the middle of the storm. With personalised news streams sharing fake news and an increasing risk of black-hat hacking, we se how A.I. is being driven by the wrong anchors.

There are specific actions to fix this trend, like adhering to a strong moral code of conduct (which embeds morality into algorithms), reducing bias from training data, and embracing privacy and data protection as an opportunity to do good.

Filip also emphasised the importance to cultivate and embracing the learnings of fields like psychology and philosophy, as this leads to realise the effects (both positive and perilous) of the way algorithms are built.

Last but not least, Filip said that events like Data Summit are the centerpiece of sharing knowledge and congregate people, and to make people reflect about how can ethics can be part in the “chain” running from an algorithm’s data source to the output. He expects that initiatives like the General Data Protection Regulation (GPDR) can spread for a positive use of data.

A takeaway from Filip’s presentation

“Researchers should be driven by curiousness, ethic and morality. Not by law or politics.”

 

We look forward to talk soon to Filip on the improvement on the use of A.I. for a better world!

Filip’s interview:

Filip’s presentation recording:

To be published soon

Filip’s deck:

Filip’s presentation sketch:

DIS17_FilipMaertens