Let me expand on this idea a bit here. I believe you cannot teach someone to become a Data Scientist but you can guide them through the process and facilitate their learning path. Remember all mistakes you have been committing when you first learned how to code and analyzed your first data set using python or R? Now you can share this knowledge with your students so they can avoid the mistakes you have done in the past.
But is this the best learning path they could take? I am going to share here a few tips that I have learned from mentoring Data Science students over the past few months.
image.png 453.8 KB
1. It is the students' objective to learn, not yours to teach them.
It is important for students to understand that to learn new concepts they are the ones that need to put primary effort. You cannot magically open someone's brain and fill it with the new information. They are the ones that need to seek the information, read data science material and practice given exercises.
This actually leads directly to the next point…
2. Advice your students to surround themselves with Data Science.
They should not only be following some exercises or tutorials outlined in a single book or course. It is important that they start reading Data Science articles from early on a variety of topics: machine learning, data cleaning, data visualization, job opportunities, basically anything connected to Data Science. It is a good idea if they subscribe to Data Science video channels and podcasts so the immersion comes from a variety of resources.
3. Leverage previous work experience or interests.
This tip is addressing especially people that are trying to transition to Data Science form a different domain. I would encourage them form early on to work on projects that are connected to the area they are currently specializing in. So if someone is working in marketing I would advise him to choose data sets for his early projects that are connected to marketing or sales. This way the learning curve becomes less stip and they can leverage their expertise so their initial projects can shine.
4. Let your students make mistakes.
I believe this is an important part and this is how learning should be done. People learn from making mistakes. Your student may use the wrong function for a particular operation. It is important to let them discover it and improve it by themselves. They may do the same mistake once or twice but the third time they will get it done correctly. This is usually more effective than telling them to avoid this mistake in the first place.
5. Teach students how to read the documentation and find information.
Teach your students how to read the documentation and how to find information. My students often asked me questions like “how do I do that in pandas?” or “what do I need to change so my graphs will look like this?”. I believe you should not answer these questions directly but refer your students to places where they can find the answer. It would normally be library documentation, particular tutorial or stack overflow discussion. With time they will learn how to get this information without you even pointing them to appropriate resources.
6. Be patient, positive and encouraging.
Last but not least remember to be patient, positive and encouraging. People learn at different rates and it may take a longer time for someone to absorb a particular piece of knowledge. In this situation, you can try to remember how hard it was for you to learn some more complex programming concepts. You can actually share this experience with your students as an example of learning something difficult and telling them how you managed the situation.
“How to teach Data Science”– Magdalena Konkiewicz Tweet