The Meaning of Data Science.
Alexander Weiss, Head of Data Science, writes about purposeful work at a Data Dive by Data Science for Social Good Berlin.
Have you ever thought about how purposeful your job is? I do from time to time. My usual train of thought looks like this: GetYourGuide is a purposeful business. We want to turn trips into amazing experiences. We are passionate about building a great business. We are devoted to delighting our customers. This is not only part of our core values – it’s something that we consider every day when prioritizing our tasks.
From a data science point of view, I disagree with Jeff Hammerbacher’s famous quote: "The best minds of my generation are thinking about how to make people click ads. That sucks." Yes, performance marketing is an important part of my daily work, but why should the task of making potential customers aware of our service be less worthwhile than, say, pointing them to the right product once they are on our website? I’ve never heard anybody saying recommender systems suck; both tasks simply optimize different parts of the customers’ journeys.
Then again, there is a spark of truth in Jeff’s quote. None of my daily tasks as a data scientist at GetYourGuide will help to solve any of this century’s most urgent problems – not in an obvious way, at least. I’m working in a purposeful business, but it’s still a business. I began to compensate for this missing part of my work with volunteering: I spent money for appropriate projects (something that we also do at GetYourGuide, such as for Refugees Welcome or GoVolunteer) and did some office work for NGOs. Still, it was not satisfying: My contribution felt small in relation to the actual problem. More than this, I was not exploiting my core skill: data science.
The situation only changed when I met like-minded data scientist Daniel Kirsch. He told me about a growing worldwide community of data scientists who were using their skills to make NGOs data-driven. NGOs often have similar problems to companies (targeting the right audience for donations, aligning supply and demand, evaluating the project’s success on an analytical basis), but they don’t have the same financial capacity to hire experts. Even worse, NGOs are often unaware of just how much value is hidden in their data. There are several initiatives worldwide that work to change this. The most well known is probably New York-based DataKind, which runs projects in several countries. Unfortunately, Germany is not one of them. Daniel Kirsch’s idea was to bring the same spirit to Germany.
Today, roughly two years after this first meeting, Data Science for Social Good Berlin (DSSG) supports German NGOs in data challenges. The preferred means of support is Data Dives: data science hackathons. Volunteers meet on a weekend to work full-time on well-defined data questions. This event is just the highlight of a longer process. NGOs need to be contacted, data problems must be identified and formulated, and the data needs to be preprocessed – maybe anonymised – and made available. This is done by a group of volunteers known as data ambassadors.
The attending NGOs at the last Data Dive in March 2017 were Deutsche Krebsgesellschaft (DKG, Germany Cancer Society) and SchulePLUS. DKG focuses on the systematic recording, analysis and provision of information in the field of oncology for practitioners, patients and interested persons (relatives etc.). This requires reviewing and categorising numerous scientific papers. Here, the Data Dive’s focus was on a proof of concept that modern machine-learning approaches can help to simplify – or even automate – this process. Besides other initiatives, SchulePLUS runs schülerpraktikum.de, a platform to connect German junior high school students looking for internships with companies offering them . The Data Dive’s scope for this project was on a better understanding of supply and demand: What are the students looking for? Are there significant mismatches to what is offered?
More than 30 data analysts and scientists worked for a whole weekend on these questions. The discussions about how to tackle the different problems with what tools started directly after the opening event on Friday evening. Saturday and part of Sunday were then devoted to the analytical work, resulting in a presentation of the outcomes. You can read about the achievements in more detail on DSSG’s blog, the short summary is: The outcome was meaningful for the NGOs, the volunteers learnt from each other about new tools and techniques, and everybody had the chance to extend his or her professional network. It was a great event, and I’m proud that GetYourGuide contributed as a sponsor to it. I’m looking forward to the next chance to applying data science in a truly purposeful way.
If you are interested in joining Alex and the Data Team, you can find all of our engineering openings here.