Data Profiling

Loading...

From the course by New York University Tandon School of Engineering

Information Visualization: Foundations

ratings

New York University Tandon School of Engineering

Information Visualization: Foundations

ratings

Course 1 of 4 in the Specialization Information Visualization

The main goal of this specialization is to provide the knowledge and practical skills necessary to develop a strong foundation on information visualization and to design and develop advanced applications for visual data analysis. This course aims at introducing fundamental knowledge for information visualization. The main goal is to provide the students with the necessary “vocabulary” to describe visualizations in a way that helps them reason about what designs are appropriate for a given problem. This module also gives a broad overview of the field of visualization, introducing its goals, methods and applications. A learner with some or no previous knowledge in Information Visualization will get a sense of what visualization is, what it is for and in how many different situations it can be applied; will practice to describe data in a way that is useful for visualization design; will familiarize with fundamental charts to talk about the concept of visual encoding and decoding.

From the lesson

Data Abstraction

Reflecting on Data7:38

What is Data Abstraction?7:51

Dataset Types: Tables and Networks5:58

Attribute Types5:06

Attribute Semantics7:28

Example for Attribute Types and Semantics5:17

Data Abstraction to Visualization8:30

Data Profiling5:06

Quick Recap2:57

Meet the Instructors

Enrico Bertini
Associate Professor
Computer Science and Engineering
Cristian Felix
PhD. Candidate
Computer Science and Engineering

As we have seen, performing data abstraction is

a very useful exercise to make decisions

about what are appropriate visual representations for your data.

Now before we conclude,

I want to make it clear that data abstraction is not

the only process or

the only method that you typically have to use in a data visualization project.

There is much more than that.

So, connected to data abstraction is the idea that

you have to become familiar with your data set.

As you have seen in the previous examples,

sometimes you just don't know what is the real meaning of a given attribute type.

Data can come from many different sources,

the meaning may actually not be evident at all.

And typically what happens is that either you need to do some research on your own

or if you have a client you will

need to talk with your client who is providing data to you,

to understand in a much more deeper sense,

what is the meaning of every single attribute.

So, this activity is often called data profiling.

I also like the term data familiarization

because you have to become familiar with your data.

And I want to make sure that you realize that in

practice what happens is that this activity

takes some time and also requires interacting with other people who have the knowledge,

a much deeper knowledge than you,

of what is the meaning of the attributes that are contained in your data set.

Before we conclude, I want to point out that

another very important activity that is typically

performed at the beginning of a data visualization project,

is checking data quality and doing what is called data wrangling.

So, let me talk about these two elements.

Data quality means that data often comes in

ways that are not necessarily precise or complete.

Right? For instance, many,

many real world data sets out there have missing values, right?

So, what do you do if some value is missing in one or more attributes?

You'll have to decide how to fix this problem.

Right? Or sometimes there are wrong values,

say you have a quantitative attribute and one single instance,

one single item as one value that is completely off the mark.

You'll have to detect this problem and you have to find a way to fix it.

Right? So, the activities that are necessary to identify these problems and also fix

these problems so that you can change that sum of

the values of your data or even transforming your data from one type to another,

something we will touch upon in the next lesson.

Yes, you'll need to perform these activities.

So before I conclude,

I want to point out that this course can't cover all these elements.

So, these elements are typically covered in

other courses that focus on data manipulation.

So, if you want to know more about data manipulation,

you want to learn more about data manipulation,

I encourage you to find other courses that focus on this specific problem.

In this course, we focus on the problem of the data abstraction because as I have shown,

it's very useful to you to decide what

visual representations are available and effective.

And in the next lesson,

I will give you a little information about what data transformations are useful.

Again, focusing on transformations that are useful to visualize data better.

But I think it's important for you to realize that this topic is much larger than that.

And once again, if you want to know more about it,

you'll have to find courses that address these problems specifically.

Explore our Catalog

Join for free and get personalized recommendations, updates and offers.

Coursera provides universal access to the world’s best education, partnering with top universities and organizations to offer courses online.

© 2018 Coursera Inc. All rights reserved.

Download on the App Store

Get it on Google Play