Video-2: Structured vs. Unstructured Data

Loading...

Из курса от партнера Rutgers the State University of New Jersey

Social Media Data Analytics

101 оценки

Rutgers the State University of New Jersey

Social Media Data Analytics

101 оценки

Learner Outcomes: After taking this course, you will be able to: - Utilize various Application Programming Interface (API) services to collect data from different social media sources such as YouTube, Twitter, and Flickr. - Process the collected data - primarily structured - using methods involving correlation, regression, and classification to derive insights about the sources and people who generated that data. - Analyze unstructured data - primarily textual comments - for sentiments expressed in them. - Use different tools for collecting, analyzing, and exploring social media data for research and development purposes. Sample Learner Story: Data analyst wanting to leverage social media data. Isabella is a Data Analyst working as a consultant for a multinational corporation. She has experience working with Web analysis tools as well as marketing data. She wants to now expand into social media arena, trying to leverage the vast amounts of data available through various social media channels. Specifically, she wants to see how their clients, partners, and competitors view their products/services and talk about them. She hopes to build a new workflow of data analytics that incorporates traditional data processing using Web and marketing tools, as well as newer methods of using social media data. Sample Job Roles requiring these skills: - Social Media Analyst - Web Analyst - Data Analyst - Marketing and Public Relations Final Project Deliverable/ Artifact: The course will have a series of small assignments or mini-projects that involve data collection, analysis, and presentation involving various social media sources using the techniques learned in the class.

Из урока

Introduction to Data Analytics

In this first unit of the course, several concepts related to social media data and data analytics are introduced. We start by first discussing two kinds of data - structured and unstructured. Then look at how structured data, the primary focus of this course, is analyzed and what one could gain by doing such analysis. Finally, we briefly cover some of the visualizations for exploring and presenting data.Make sure to go through the material for this unit in the sequence it's provided. First, watch the four short videos, then take the practice test, followed by the two quizzes. Finally, read the documents about installation and configuration of Python and R. This is very important - before proceeding to the next units, make sure you have installed necessary tools, and also learned how to install new packages/libraries for them. The course expects students to have programming experience in Python and R.

Video-1: Introduction3:02

Video-2: Structured vs. Unstructured Data10:34

Video-3: Analyzing Structured Data10:31

Video-4: Visualization of Data8:40

Познакомьтесь с преподавателями

Chirag Shah
Associate Professor
Information and Computer Science

Hello, in this sub unit, we'll talk about data, we'll talk about

what's the difference between structured data and unstructured data.

So what is data?

Well, we see it all the time, we've seen it and all kinds of websites,

mobile devices, on our computer, or even outside of online world.

Here's one example.

Here's data that's, it's a GDP data from World Bank.

And I'm sure you've seen something like this, where there's a spreadsheet,

with some country information,

and here is the columns and then there is a GDP.

So this is one kind of data, which is quite common, we encounter in many places.

Here is another example and again this is a spreadsheet where this is

as a T-data from a New York City open data site and

again you can see different school names here as at this course and so on.

This is an example of what's called structure data because here,

all the data fields and this case such as a school name,

the SAT score and, it's broken down into reading, math, writing scores.

So those things, those numbers are marked, there are these

fields that are marked, and so this is an example of structure data.

Here's another example now.

This is a blog example where there are some things that

are defined, but most things here are what's called unstructures,

they may appear to be organized at least on the layout, but

the fields are not marked, they're not presented in a way that

shows predefined kind of a data.

This typically happens with text, I saw here.

Can see that in the blog there was likely to be text, but

also has a combination of text, and images, and so on.

And I'm sure you've seen something like this if you have spent any time on

the Internet.

Something like YouTube, where it's multimedia data,

and we say it's multimedia because it's primarily video but

there is also mix of text and often images, and

other source of data that's combined in this one place.

So there are different formats here.

There are different structure here.

Some of that is structured data, the things that are marked.

And other are unstructure because there're not well defined,

they are free fine text and tags and other things.

And here is an example from another social media site that you

may have heard about Twitter.

So again, here some things are marked clearly some things are not.

So it's a mix of all different formats and

data so there's the basic unit, a tweet is text but then

that also includes lings and often know, videos, and images, things like that.

And so, some of the things are structured, some others are not.

So of course, now we've seen these examples, we can provide a more formal

definition, and understanding of what the structured and unstructured data are.

So structured data is where you have some kind of a high level of organization.

And then, this typically happens in relational databases,

spreadsheets which also example short sheets.

So that's what defines some structured data.

Here's an example from,

it's a data schema or data model,

from my SQL which is structured query language.

And you can see here these are, and it's okay if you don't know

SQL or never used it, but, essentially, these are different tables, so

there's departments, employees, salaries, titles, and then so on.

And for each of those tables there are some

characteristics of the data that would go into it.

So for instance, when you're talking about an employee, there is going to be their

birthday, their first name, their last name, gender, hire date, things like that.

So these are defined labels.

In order to describe an employee,

you have a predefined set of labels that goes into there.

So the structure

data will depend on a data model that's typically already defined.

And that definition, that data model tells us how the data about,

for example an employee, should be stored presented.

And so, the nice thing is because there is a predefined data model

where somebody tells us how to store the data,

it's easy to enter that data, it's easy to store it, and

it's also easy to retrieve from it because now you can, ask a very specific question.

For instance, you can ask, give me the record of an employee named Jane Doe.

And so, you're able to say first name Jane, last name Doe, from that employee

table, because those fields, first name, last name, are defined.

So for instance, if you want to find out how many

employees will be retiring in the next five years,

you could query this employee table using the birth

date to find out their age and use that information.

So you're able to specifically ask for a field because those fields are defined.

So that's the nice thing about structured data that since the data model is defined,

a data model is present.

All those fields are define and easily, can be easily restored, queried,

retrieved, analyzed.

Unstructured data on the other hand those not come with those kind of labels,

does not come with the predefined data model.

And so, here's an example.

It's a Wikipedia article, or as a screenshot of Wikipedia.

But a lot of things on Wikipedia, those articles, they have free flowing text.

Now, there could be mark up text that defines a formatting of the tags.

It could be images, videos, links.

But those things are not going into specific predefined data model categories.

This is just what we saw before in the employee record.

And so, photos, graphic images, the videos,

PD apps, emails, all those things

tend to fall in unstructured data because the data has been generated, stored.

But they're none going into specific data fills.

So that sort of the unstructured data.

When it comes to social media, we see mix of both.

So we see things like this where here's an example where you

collected this data from Brexit.

So based on the hashtags that are used, so

these data that are basically tweets and we can represent them using

this kind of numbers and say, well, how many tweets are using #brexit.

How many tweets are using #voteleave and so on.

And so, this is an example of structured data because we are looking for

a specific, we are able to look for a specific thing and count them.

And we are able to look for

specific things because they are in the structure format.

Then there is unstructured of course.

There is a lot of text and so when you're looking at a lot of text,

which is not clearly labelled.

You need to do different kind of analysis,

you need to do different kind of representation.

So here's an example of text where you can do sentiment analysis and

this is where, you have free flowing text.

It's not defined if something is somebody's name.

If something is a place, a time, it's just text.

And so, we can't simply count or compare.

So when it comes to something like sentiment analysis,

that's down on this kind of unstructured data.

So typically, structured data helps us do a lot of number-driven

quantitative analysis, unstructured data we end up doing,

you can still do quantitative analysis,

there is also a lot of qualitative questions that you can ask.

In this course, we'll be focusing on structured data,

most times the data that will derive will be of that nature,

and so we'll be doing mostly quantitative analysis.

So what we did in the subunit, we talked about structure data,

which is where the data fields are defined,

and the data is organized according to those fields, or labels.

So there is a preexisting data model, and

unstructured data is when this kind of data model does not exist.

There are no predefined organizing scheme.

So that's the main difference.

And in this course, we'll be mostly dealing with structured data.

So that's the end of this subunit.

Ознакомьтесь с нашим каталогом

Присоединяйтесь бесплатно и получайте персонализированные рекомендации, обновления и предложения.

Coursera делает лучшее в мире образование доступным каждому, предлагая онлайн-курсы от ведущих университетов и организаций.

© Coursera Inc., 2018 Все права защищены.

Загрузить из App Store

Загрузить в Google Play