Data obfuscation Second Draft

[Tony911029]

1. Fine-tune distribution to be more realistic
2. Make normal distribution have a mean of 0
3. Simplify shifting mechanism
4. Comment and refactoring codes
5. Strip out some new line character for some files (for some reason, some csv file name has a new line char at the end)

Randomly generate 50 sample curves for each distributions:

Side note:
Being an early logger doesn't mean the logger will always log early. They are just more likely to log early than late.