Also crucial: Have a development process - and stick to it! A question you can always ask is: What if I gave you the perfect blackbox? What business actions are implied and what processes would trigger the results? "Your job as a data scientist is to ask these questions - and much more, just never stop asking questions", says von Rohr. 

But never forget: resist the urge to publish everything to early. An angry e-mail? Write it - but don't send it immediately. Results which are not ready can always backfire, so wait some time before you publish them. 

Dr. Diana Mantel

"Treat your project like a startup", recommends von Rohr. Don't be afraid to pivot, you may have simply misunderstood the requirements. For example take a product you already have and make another product out of it. 

What is also important: Have an elevator pitch. Be always ready to convey the value proposition of your project in 30 seconds. "That gives you a clear insight what you project is about, for whom it is made and how it will work out", says von Rohr. Even if you never meet somebody important in an elevator in your life: you have a clear idea of your project by yourself.

Dr. Diana Mantel