Getting insights from data is the new default, and we even have a fancy name for it: Data Science. It has been like this for a while now. However, new techniques are now more accessible thanks to hardware improvements and great libraries.
There’s a big secret that rookies and veterans alike tend to overlook. Ok, it isn’t a very big secret but one of the most often overlooked. In the same way that the secret for an awesome body involves a lot of consistent boring work.
I’ve always known this the sh!t machine, but to avoid writing a page full of sh!t, I’m going to go all Randian and call it the “check your premises method”.
This is based on the objectivist philosophy developed by Ayan Rand and captured by:
Contradictions do not exist. Whenever you think you are facing a contradiction, check your premises. You will find that one of them is wrong.
Ayn Rand , Atlas Shrugged
Check your premises
If you ask dumb questions, you’ll get dumb results.
If your input is bad, your output is going to be bad.
If you don’t know where you are going, you’ll end up somewhere else.
You get the gist. All the fancy algorithms in the world won’t save a bad premise.
Scientific method
Data scientists work across industries surrounded by people that haven’t devoted enough time understanding the ins and outs of Data Science. It’s up to them to guide the stakeholder, set expectations, and follow through with the scientific method that puts the name Science to this underappreciated profession.
It’s basically the following:
- Help to formulate quality questions.
- Ensure that the data is of the highest quality possible and the best available.
- Set hypotheses.
- Build a model (or as many as you need).
- Test.
- Analyze and try to answer the question.
Step number 2 is invented. Take your time making sure that you have what you need before attempting to set hypotheses and building a model.
Otherwise, you fall prey to the too-common pitfall: perfect useless models.
Return on investment focus (ROI)
A good data scientist is the one that allows time to step back and think about the ROI (return on investment), of the project. On top of that, the good data scientist tries his best to work from sound premises, ie good questions, and good data.
The black box doesn’t do miracles. You’ll end up with a function of what you start with.
If you found this article interesting let me know in the comment section below! If you would like to have me ramble about data science, games, product management feel free to get in touch.
Additional resources:
Scientific method https://en.wikipedia.org/wiki/Scientific_method
Check your premises quote from Ayan Rand https://www.brainyquote.com/quotes/ayn_rand_163204