Handling Negative Result in Data Science Project!
You ran your analysis or experiment and the results are negative? Great 👍! This is a normal part of your data science experiment and in this video I'll show you why you should learn to love them!
Note: This video sounds more like a therapy session than a data science tips haha
Table of Content
Introduction: 0:00
Incorrect Hypothesis : 2:20
Technical Errors : 4:00
Contradicting Literature: 10:38
Conclusion: 12:10
There are three main reasons why your results are negative (i.e. not what you expected). Each of these type of negative results have tremendous benefit for your research project as a whole and you should always strive to understand in which bucket your current result is falling into.
1. The original hypothesis was inaccurate and based on false and incorrect assumptions.
2. Technical errors, which include choosing a misfit study design, use of inappropriate statistical methods or plain-old bug in your analysis.
3. Your result contradict an earlier published reports
Whatever is the type of your negative result, take it with a big spoonful of positivity. Figure out in which of the three buckets it lies and then iteratively make your analysis better so that you get to that sweet positive result!
It will not only makes your analysis better, but you as a data scientist will grow from the experience. You will be able to learn new tools, software engineering practice and process to make your analysis rock-solid!
----
Join the weekly AI newsletter: https://yacinemahdid.com
Join the Discord for general discussion: / discord
----
Follow Me Online Here:
LinkedIn: / yacinemahdid
GitHub: https://github.com/yacineMahdid
Twitter: / yacineaxya
Instagram: / yacine_mahdid
___
Have a great week! 👋