“What challenges are you tackling at the moment?” I asked. “Well,” the ex-academic said, “It looks like I’ve been hired as Chief Data Scientist… at a company that has no data.”
I don’t know whether to laugh or to cry. You’d think it would be obvious, but data science doesn’t make any sense without data. Alas, this is not an isolated incident.
Data science doesn’t make any sense without data.
So, let me go ahead and say what so many ambitious data scientists (and their would-be employers) really seem to need to hear.
If data science is the discipline of making data useful, then you can think of data engineering as the discipline of making data usable. Data engineers are the heroes who provide behind-the-scenes infrastructure support that makes machine logs and colossal data stores compatible with data science toolkits. …
Attracted by the lure of lucrative jobs, these hucksters give legitimate data professionals a bad name.
[In a hurry? Scroll down for a quick summary at the bottom.]
Chances are that your organization has been harboring these fakers for years, but the good news is that they’re easy to identify if you know what to look for.
Data charlatans are so good at hiding in plain sight that you might even be one without even realizing it. Uh-oh!
The curse of dimensionality! What on earth is that? Besides being a prime example of shock-and-awe names in machine learning jargon (which often sound far fancier than they are), it’s a reference to the effect that adding more features has on your dataset. In a nutshell, the curse of dimensionality is all about loneliness.
In a nutshell, the curse of dimensionality is all about loneliness.
Before I explain myself, let’s get some basic jargon out of the way. What’s a feature? It’s the machine learning word for what other disciplines might call a predictor / (independent) variable / attribute / signal. Information about each datapoint, in other words. …
Instead, here are some more colorful candidate names for your amusement.
Painful value: They make you calculate it in class without explaining it to you properly; no wonder your brain is hurting. Honorable submissions in this category also include puzzling value, perplexing value, and punishing value.
Pesky value / problematic value: Statisticians are so tired of seeing ignoramuses abuse the p-value that some of them want to see it abolished. …
Before we dissect the nature of analytical excellence, let’s start with a quick summary of three common misconceptions about analytics from Part 1:
While the tools and equations they use are similar, analysts and statisticians are trained to do very different jobs:
If you’d like to learn more about these professions, check out my article “Can analysts and statisticians get along?”
Analytics is not marketing. …
In a nutshell: A/B testing is all about studying causality by creating believable clones — two identical items (or, more typically, two statistically identical groups) — and then seeing the effects of treating them differently.
Scientific, controlled experiments are incredible tools; they give you permission to talk about what causes what. Without them, all you have is correlation, which is often unhelpful for decision-making.
Experiments are your license to use the word “because” in polite conversation.
Unfortunately, it’s fairly common to see folks deluding themselves about the quality of their inferences, claiming the benefits of scientific experimentation without having done a proper experiment. …
Experiments allow you to talk about cause and effect. Without them, all you have is correlation. What is correlation?
IT’S NOT CAUSATION. (!!!!!)
Sure, you’ve probably already heard us statisticians yelling that at you. But what is correlation? It’s when the variables in a dataset look like they’re moving together in some way.
For example, “when X is higher, Y tends to be higher” (this is called positive correlation) or “when X is higher, Y tends to be lower” (this is called negative correlation).
If you’re looking for the formula for (population) correlation, your friend Wikipedia has everything you need. But if you wanted that, why didn’t you go there straight away? Why are you here? Ah, you want the intuitive explanation? Cool. …
In my previous article, I explained why you shouldn’t look to statistical inference for truth. Given the prevalence of statistical techniques in scientific research, what does this mean for science?
(For those who insist that you need credentials to have an opinion about science, this jerk of an author holds graduate degrees in neuroscience and mathematical statistics. Glad we got that out of the way.)
A hypothesis is a description or explanation, but it needn’t be true. If it amuses me, I can hypothesize that no human is taller than five feet. …
INFERENCE = DATA + ASSUMPTIONS. In other words, statistics does not give you truth.
Here are some standard misconceptions:
They sound like fairytales, don’t they? That’s because they are!
There is no magic in the world that lets you make something out of nothing, so abandon that hope now. That’s not what statistics is about. Take it from a statistician. (As a bonus, this article might save you from wasting a decade of your life studying the dark arts of statistics to chase that elusive dream.) …
Imagine that you’ve just managed to get your hands on a dataset from a clinical trial. Exciting! To help you get in character, I made up some data for you to look at:
Pretend that these datapoints map out the relationship between the treatment day (input “feature”) and the correct dosage of some miracle cure in milligrams (output “prediction”) that a patient should receive for over the course of 60 days.
(1,28) (2,17) (3,92) (4,41) (5,9) (6,87) (7,54) (8,3) (9,78) (10,67) (11,1) (12,67) (13,78) (14,3) (15,55) (16,86) (17,8) (18,42) (19,92) (20,17) (21,29) (22,94) (23,28) (24,18) (25,93) (26,40) (27,9) (28,87) (29,53) (30,3) (31,79) (32,66) (33,1) (34,68) (35,77) (36,3) (37,56) (38,86) (39,8) (40,43) (41,92) (42,16) (43,30) (44,94) (45,27) (46,19) (47,93) (48,39) (49,10) (50,88) (51,53) (52,4) (53,80) (54,65) (55,1) (56,69) (57,77) (58,3) (59,57) (60,86) ... …