The Ghost of the Future of Data Science

Let’s aim to build solutions for real-world problems

At the end of the year I like to reflect on all that I have learned and project it to what I want to do next year. Predicting outcomes of very complex matters is tantamount to sacrilege in our field, and rightly so. But as data science is still an emerging field of research, without an agreed upon definition, it’s also still the sum of all the work of colleague researchers in the field. Of which a bunch of them work at Jheronimus Academy of Data Science (JADS).

The following is an outline of things that I will be busy with next year at JADS. It’s by no means a clear line of reasoning (I wish!) but I hope that you more or less understand what I am getting at.

1. Data scientists engineer solutions to real-world problems

Earlier this year Duncan Watts has raised the important question if social science should be more solution-oriented. I would like to ask the same question to data science researchers around the world.

Industry has been a major inspiration for the success of the field of data science, and in that sense it has a very applied quality to it. But I work in the realm of fundamental research, and if data science wants to have any future at all, it needs to have a fundamental part as well.

In my vision (which takes the ideas of Duncan Watts’ paper in Nature on solution oriented science as starting point) there is a place for fundamental research where the data scientist builds or uses a system that solves a real-world problem, and from the data that is generated we can both measure the effect and give feedback to exisiting theory or develop new.

So instead of always using existing datasets from existing systems, which is often the case in industry, we are more focused on generating new data from new systems especially build to solve problems that we are researching.

Also, although it is great that data science, or maybe more specifically big data research methods, can be a really good new tool for many disciplines, including social science, one could argue that it’s not the mission of data science to improve other disciplines. In this vision it’s the responsibility of the academic disciplines themselves to innovate.

2. Why data science can’t do without design (or entrepreneurship)

As showcase research project I took a personal observation, the low quality of wellbeing measurement at elementary schools. I have a son of 7 and a daughter of 4. Every morning when my daughter enters the classroom she has to score her happiness on a scale of ☹️-😐-😊. Her teacher than has an idea how the children are feeling, this kind of self-reporting is actually very common in wellbeing research. But no data scientist I know is content with doing this kind of subjective research.

So I started a project called Sensing Happiness with my colleague Martin Atzmüller. The goal of this project is:

How can we sense, predict and influence happiness?

1. develop a more objective measure of happiness using social sensors
2. develop predictive models based on social interaction data
3. develop a wearable that gives personalized interventions to inspire happiness

The system here is a wearable, that both gathers unique data on children’s wellbeing, that data can be used to build predictive models and the wearable is also used as intervention or persuasive technology.

The persuasive effect of influencing additional happiness, should be the start from where we begin with theorizing.

Because we are focused on real-world or real-life problems, the wearable should be user-friendly and probably playful given the context. I wouldn’t pick a random statistician to build this wearable, or any data scientist for that matter. So I think that we have a natural urge to work together with designers on intelligent user systems.

Less clear is the link between entrepreneurship and fundamental data science. That’s a discussion I am already having with my colleague Mark van de Pol 🕶👟👣. Within this view of data science I believe good science means good bussines as well, when it’s your business to solve relevant societal problems of course.

3. Data science, as an academic field, shouldn’t cater to corporate agendas. It should fix society.

As I wrote earlier this year, in data science should monitor big brother, I think we should not follow the money and cater to corporate agendas. If we do this, we might just end up on the wrong side of history.

The last months we have witnessed a significant change in the public understanding of the impact of large tech companies on our society. Just yesterday we discussed the awesome rant by Scott Galloway in our monthly research meeting.

In the many discussions I have had this year it became clear to me that academics, industry leaders and the public are pretty overhelmed by the speed of technological developments.

But next to that, many of us don’t know whether this technological progress is also serving human progress. Industry has let the genie out of the bottle to paraphrase Josh Elman, and doesn’t seem to be able or willing to put it back. Regulators are struggling as well. And our field hasn’t yet stood up as a whole to show the moral capacity and lead us to a brighter future. Would be great if this happens soon!

Some interesting observations by Josh Elman:

"Facebook can’t put the genie back in the bottle. Social media has transformed a lot of things - in many good ways and in many negative ways that none of us involved anticipated at the time."

"I still like Ev Williams’s phrase on this the best: if the world was data driven, it would see that people love looking at car accidents and create more of them. It is time for FB and Twitter to move beyond car accidents."

But to me it’s evident that the data science research community has to develop a clear vision on what role they will play in the academic community and in society at large. Again my current thinking goes as follows:

1. Data scientists engineer solutions to real-world problems
2. Data science can’t do without design (or entrepreneurship)
3. Data science should help fix society

I don’t dare to predict what the future will bring. But I will hopefully work on some projects that aim to fix the future. If anything, one of my main projects next year is called FixTheFuture… more on that in 2018.

See you at the other side.