Become Sherlock Homes by applying Feature Engineering in practice
- Basic statistics and mathematics is your compass in the field of Data Science. You must have it to have a sense of direction.
- You are better-off choosing one of R or Python or Matlab or any other platform, and become a master in it than attempting to get dirty with all of these.
- Right from the start, version-control your work with a tool like Git/Github/Bitbucket. It can save you from a lot of headaches.
- Don’t under-estimate the importance of understanding the problem domain. Domain knowledge is the secret weapon for Data-Manipulation.
- Don’t under-estimate data cleansing. It pays to cleanse your data.
- Pay attention to how you impute your data and document why you took that approach.
- Don’t ignore exploratory data analysis (a.k.a EDA). It not only improves your problem-domain knowledge, but also instigates your creative juices towards solutioning.
- The EDA that you do in the beginning is the road that you lay for your presentation on insights and solutioning to business stake-holders in the end.
- Master the art of Feature Engineering to improve your probabilities of becoming the celebrity Sherlok Homes of Data Science.
- Don’t jump into esoteric modelling. Start simple.
- Don’t jump into a model because it is popular. Know how the model works and how you can tune it to improvise it.
- It is better to make Data Scaling as mandatory part of your data pipeline, than have it as an optional thing. It gives more options to try out various models.
- Data-modelling is both the art and science of finding that optimal trade-off between bias and variance. Enjoy the game.
- Almost always you don’t get a 100% accuracy with your data-modelling. And when you get it, question yourself and your approach. Triple-check your understanding and get it reviewed before celebrating. Save yourself the unwarranted heart-break.
- Find opportunities to pair-up with someone to teach or learn. You will end up learning something for sure. It has its definitive ROI for you.
- Up-skill with deliberate practice. How well you practice (quality)is equally important if not more than how much you practice (quantity).
That is all to the list…for now.
Got something to add to the list? Do share your pro-tips as comments.
Like what you read? Don’t forget to click the clap button and share it with your circles.
Have a feedback or two for me? Do drop your comments below. I’m all ears for it.