We recently read that The New York Times called data cleaning janitor work and said that 80 percent of a data scientist's time will be spent doing this kind of cleaning. As we can see in the following figure, despite its importance, data cleaning has not really captured the public imagination in the same way as big data, data mining, or machine learning:

Who can blame us for not wanting to gather in droves to talk about how fun and super-cool janitor work is? Well, unfortunately—and this is true for actual housekeeping chores as well—we would all be a lot better off if we just got the job done rather than ignoring it, complaining about it, and giving it various demeaning names.

Not convinced yet? Consider a different metaphor instead, you are not a data janitor; you are a data chef. Imagine you have been handed a market basket overflowing with the most gorgeous heirloom vegetables you have ever seen, each one handpicked at the peak of freshness and sustainably produced on an organic farm. The tomatoes are perfectly succulent, the lettuce is crisp, and the peppers are bright and firm. You are excited to begin cooking, but you look around and the kitchen is filthy, the pots and pans have baked-on, caked-on who-knows-what, and, as for tools, you have nothing but a rusty knife and a soggy towel. The sink is broken and you just saw a beetle crawl out from under that formerly beautiful lettuce.

Even a beginner chef knows you should not cook in a place like this. At the very least, you will destroy that perfectly good delicious basket of goodies you have been given. And at worst, you will make people sick. Plus, cooking like this is not even fun, and it would take all day to chop the veggies with an old rusty knife.

Just as you would in a kitchen, it's definitely worth spending time cleaning and preparing your data science workspace, your tools, and your raw materials upfront. The old computer programming adage from the 1960s—garbage in, garbage out—is also true with data science.