What are Data and Data Literacy: Study Hall Data Literacy #1: ASU + Crash Course

Data science.

Data analytics.

Real-time data.

Big data.

Little data?

It can feel like no matter what we’re doing — working or studying or just surfing the

internet — we’re being bombarded with a deluge of data buzzwords.

With all that buzz, it sounds like there’s data everywhere!

But what really is data?

And is it really everywhere?

In fact, without even realizing it we’re constantly producing and consuming data.

Even right this second!

Watching this video gives YouTube data on what to recommend to you in the future, like

more Study Hall videos!

So, as we swim through the sea of statistics, analytics, and visualizations, we have to

think critically about all the data coming at us.

Some of it is helping us explain the world around us, and some of it is being used to

stretch the truth.

All data has a story to tell — but should we believe it?

Together, we’ll learn how to decide.

I’m Jessica Pucci and welcome to Study Hall: Data Literacy presented by Arizona State University

and Crash Course.

Sure, we hear about tech and social media giants like Google or NASA or Facebook wielding

data to change the world, but what about little ol’ you and me?

Do we really use data?

Well, every day we’re observing, collecting, and measuring information.

Take something basic, like the weather.

When packing for a trip or heading out for the day, ideally we’d think about the whole

range of temperatures we’ll see, not just something in the middle.

I mean, I’m hot then I’m cold.

I’m shivering then I’m sweating.

Or if it’s spring, maybe we’re weighing the likelihood it’ll rain versus having

to drag around a raincoat and umbrella.

Each individual piece of information -- like the temperature at exactly 4 pm yesterday

or the total amount of rain London got last month — is a single data point, or datum.

But all of that information put together — like the full temperature range for our trip or

the entire rain forecast for the day — is data, or specific information we collect

to make decisions.

You’ll hear me say data...but you’ll probably also hear me say data.

Both are fine, so pick your favorite.

And if we read up on the weather across all the cities we plan on visiting on our grand

tour, we’d have a data set, or related collections of information.

Now that might sound like we’re compiling pages of spreadsheets before heading out on

a weekend getaway.

And if that’s your vibe, you do you!

But no matter how we’re processing all the information coming at us — tracked in a spreadsheet

or just informally in our heads — our ability to read, analyze, create, and talk about data

is our data literacy.

And it’s what we’ll work on throughout this course as we learn how to decide what

story data is telling us, and whether we should believe it.

The first step to building our data literacy is knowing what data is made of.

We can think of data as being broken up into variables, or particular types of information

that we can use to split up our larger data set.

To take a closer look at what variables are, let's get to the data point!

Suppose you were out hiking and you accidentally fell and broke your arm.

Luckily, your friends are there to take you to the emergency room...and help you fill

out the mountain of paperwork the nurse hands you.

Each answer you scribble down gives the doctor data they can use to decide how to help heal

your arm.

All the information together makes up a data set on you and your medical history.

Now some questions, like “Age?” or “Height?”

give the doctor quantitative data, or data that’s represented by numbers, or quantities.

But not everything about us can be written as a number.

Like the question: how the heck did you break your arm?

Information that’s describing qualities something has or a category it belongs to

is called qualitative data.

Other questions on the form can be answered more descriptively,

like “Have you ever broken a bone before?”, “Which hand do you write with?” or “Have

you ever been here before?”

And that’s still data!

Qualitative data is often represented by text, like the word “collarbone” if you broke

your collarbone falling off a swing when you were 6.

Or we can use a single character to save space, like L if you’re left-handed.

And in special cases, the data might be a boolean — data that can only be either true

or false — and you’d answer “Y” for “yes” or “N” for “no.”

And after you make it through all the paperwork, the doctor might add data in other ways.

If they take an x-ray, this photo becomes a data point in your medical history.

Or they might record an audio clip of your heartbeat or video of an ultrasound.

Your file has a long story to tell.

Now, can I sign your cast?

We definitely provide a lot of different types of information to doctors every time we visit

them, and our medical histories can get pretty large over time.

Particularly if you’re a klutzy adventure lover!

But is that “big data”?

The line between data and “Big Data” status is kinda blurry, but big data is often characterized

by the 3 Vs: volume, variety, and velocity.

Your medical paperwork alone might not be voluminous, but consider all of the medical

forms filled out across the globe.

That is a lot of paper!

Big data takes up a lot of virtual or physical space with its high volume.

And medical data certainly contains a variety of data types from numbers to text to photos.

Basically all that qualitative and quantitative stuff we were talking about.

Big data contains a large variety of information.

Finally, the velocity of medical data is how fast it grows over time.

More people means more histories, and as it becomes easier to collect and store this information,

data grows and grows.

Big data is speeding down the analytics highway.

Metaphorically.

From the sound of it, big data is unwieldy to wrangle.

But if everybody kept their personal arm-injury-data to themselves, it would be harder to train

medical people to diagnose and treat all kinds of things!

Being able to navigate a dataset, even if it’s big, is important to data literacy.

With the background to ask the right questions, we can better uncover and understand patterns

that can be hidden in a treasure trove of data.

Without our help and interpretation, “big data” is powerless.

And because of how many people contribute to a big data set, it’s important to our

data literacy to understand what it really is, why it’s so valuable, and especially

the ways it can be flawed.

But we don't always need big data to answer questions that interest us.

Sometimes, just a little data will do!

For example, maybe we want to order a cake for our birthday, but we aren’t sure which

of the two new bakeries down the block are worth a try.

This is important.

There’s cake at stake!!

Information — or data — on what other people think can help us judge the deliciousness.

We better hit the streets and find out the truth from their customer populations, or

the complete collection of people or items we’re considering.

If we’re super ambitious, we’d get every customer each bakery has ever had to rate

their experience.

In the end, we’d have a census — or information from every member of a population — of each

bakery’s customers.

But that’s no small feat.

In practice it’s probably too hard, too expensive, or too time-consuming to ask everyone

for their opinions.

Instead we can ask a smaller sub-collection, or sample, of the population.

If we pick them the right way, the hope is that the ratings in the sample are representative

or have the same patterns and trends of ratings from the entire population of customers.

Luckily, there’s a sample of 0 to 5 star ratings already available online.

To make a data literate cake decision, we’ll look for the bakery that gets higher ratings

most consistently, and there are a couple of ways to think about that.

The distribution of the ratings tells us how many of each rating exists in our data and

how they’re spread out.

It’s helpful to notice how many 5’s there are versus 0's, but we need to summarize the

distribution of ratings into a few numbers to make a more informed decision.

We can start by thinking about the “middle” of the rating distribution, or the central

tendency of the data, with three metrics.

Hopefully this represents what a typical customer thinks of each bakery.

First, the mean or average is the sum of all of the ratings divided by the total number

of ratings.

Second, the median is the rating in the middle of the list if we lined up all the ratings

in order from smallest to largest.

And third, the mode is the most frequent rating.

All these ways of determining the distribution’s center have their challenges though.

Like, if three people gave 0 stars but four people gave 5 stars, we’d say the mode is

5 stars.

But the mode might not be a good representation because it’s hiding all those unsatisfied customers.

This is what’s happening with the first bakery’s ratings.

It’s possible this one unhappy customer is an outlier — or a data point that is significantly

different than the rest.

The outlier doesn't really represent how almost all of the other customers feel.

One way we can figure out if that zero really is an outlier is to also look at the qualitative data,

not just the quantitative ratings.

Reviews often come with text we can read to learn more.

Something short and angry like “I HATE THIS PLACE!” isn’t helpful and since everyone

else gave the bakery 5 stars, we can probably ignore the rating when we decide.

But something like:

“Although the staff was friendly, we were disappointed in the dryness of the cake in

addition to a mistake in the type of frosting we ordered.”

Now that’s good information.

Nobody wants dry cake!

Although it’s a little harder to manually analyze qualitative data, it often adds important

context to quantitative data.

I think we’re ready for some sprinkles!

This cupcake is great, but if our sample had only included a few ratings to start,

all of these metrics, including when we considered the review text, could’ve been unreliable.

A sample is smaller and easier to work with than the whole population, but especially

if we only have a smattering of ratings to work with, we can’t be 100% sure our measures

of center are exactly what we’d find if we asked everyone.

Basically, when we're just working with a little bit of data, we may actually need more

data to be more sure of our conclusions.

One way to check if more data would be useful is looking at the margin of error, or a window

around what we think is the center that shows how far off we might actually be.

It’s like saying, we’re working with a sample so we’re not completely sure we found

the exact number, but here’s the interval we’re pretty sure it’s in.

So a big window means we’re less confident we know exactly where the center is.

Like for our bakery ratings, if we think the center is 3, but our margin of error is the

interval 0 to 5, that’s the whole range of ratings!

We’re basically saying anything is possible.

But a small margin of error says that we’re likely on the right track.

So data is really information, and data comes up a lot because we humans rely on information.

Today we learned what data is, what forms it can take, and how to start interpreting

data to make our own decisions or double check the decisions of other people.

Next time, we’ll talk about common misconceptions in data analysis so that we can avoid these pitfalls.

Thanks for watching Study Hall: Data Literacy, which is produced by Arizona State University

and the Crash Course team at Complexly.

If you liked this video and want to keep learning with us here in Study Hall, be sure to subscribe.

You can learn more about ASU and the videos produced by Crash Course in the links in the description.

See you next time!