Fact Check and Document All Data Science Assumptions ✅

In all data science projects there is a big foundation of assumptions that are put in place in order to get the project started. Document all of them and fact-check them! fact-check
Table of Content
- Introduction: 0:00 - 1:23
- Types of assumptions: 1:23 - 3:40
- Benefits: 3:40 - 7:17
- Exemple: 7:17 - 8:40
- Conclusion: 8:40 - 10:17
Sometimes, underlying assumptions don't hold anymore with the new findings you gleaned out of some exploratory data analysis meaning that if you don't know about some assumption, you might work your way through a scientific dead-end.
One very easy way to document all assumptions when you don’t know if something is a fact or not is just to list all facts you have that relate to your project. Let’s take two examples from one study I did back in my Ph.D. days.
[Fact] All participants going into anesthesia for the EEG recording have the headset set in the same way.
[Fact] All participants have been given a heavy dose of anesthesia to make them unconscious without a doubt.
The first fact is actually an assumption I’ve unconsciously made early on in my project. When I spent more time in data collection, I’ve realized that this is absolutely not the case! Some participants who were clinically unresponsive for the study had various types of damage to their heads which made the recording a big challenge. Some even had a whole bone flap completely removed, which altered the data in a drastic way!
The second fact is technically true, all participants had heavy doses of anesthesia. However, the latter part “unconscious without a doubt” is not a fact. It’s an assumption that is hard to verify. When you look at time series recording of people under anesthesia and plot their brain activity it’s actually very dynamic. After talking to a few anesthesiologists that were helping me in the study, I’ve also realized that even under heavy anesthesia it’s possible for a patient to wake up during a surgery. So while technically a fact, it’s better categorized as an assumption, however, this is the kind of information I would just put in the limitation of a study instead of trying to solve it on top of the rest of my analysis!
----
Applying this layer of documentation in your work will instantaneously make you a much more effective data scientist! Furthermore, it will make your results interesting to debate and discuss since you are well aware of the underlying assumptions holding everything together.
Validating or disproving these assumptions is what leads to the subsequent data science project and continues the cycle of research towards more discovery!
----
Join the Discord for general discussion: / discord
----
Follow Me Online Here:
Twitter: / yacineaxya
GitHub: github.com/yacineMahdid
LinkedIn: / yacinemahdid
Instagram: / yacine_mahdid
___
Have a great week! 👋