I read this paper and understood it very well, but I miss negative impact that will happen if we don’t fix the 5 types of messy data introduced in the paper (page 5), or even how fixing them will make the analysis easier. Can any one give examples for this?
The general idea is to standardize the format of the data so that it can be used consistently across a wide range of methods for analysis and prediction.
The negative impact is the same for all 5 types of messy data: one would spend much more time implementing functions which do nothing more than converting from one very specific format to another. Additionally through these cumbersome conversions one is more likely to introduce errors or omissions in the data.
Basically following the "tidy data philosophy" can save a lot of time and makes your data readily usable with a wider range of methods.