Ask questions, lots of them
You are there to learn, you are new, people get it. You are totally fine asking a newbie question on Day 1, you probably don’t want to be asking that question after a year. Oh, and take notes, don’t ask the same question again and again.
Perfect the work others find boring
Cleaning datasets is fun! Honestly, most people don’t like it at all, but you get to write cool functions in R/ Python and do some cool plots. It is really fun.
Understand the data science life cycle
A great example of a conceptual framework appears in the Microsoft Azure documentation the “Team Data Science Process Lifecycle”:
Definitely worth your time checking out!
Learn to talk to people and collaborate
Explaining difficult concepts to people outside the data science team. So much of this game is about communication, mostly you will find data scientists work in a flat structure, so there is plenty of opportunity to collaborate.
Planning and thinking time
The more you can think, plan and understand at the beginning of a project the more you will save yourself pain down the line. You aren’t being paid for how many lines of code you can write, you are actually being paid to think, plan and solve problems. You need to ask questions, clarify and scope out the problem before you start work.
Prototype and sense checks
Most data science teams shipping data products into a production environment will be working in an Agile environment. So the emphasis will be on rapid cycles of development, iteration and prototypes. This is actually a good thing. You don’t want the situation where you go away for a year and build something the business doesn’t want. It is much better to find out early.
If you want to make sure the project you are on has the potential to lift some metric a combination of very quick and dirty ½ day models, and back of the envelope calculations can quickly let you know if the project is a good use of your time or not.
If you want more information, check out
"The Data Scientist's Journey. The Guide for Aspiring Data Scientist".
This book contains over 220 pages of data science goodness that will save you time, pain and money from a guy who has really been there, done that.