Data Governance and Profiling

Loading...

来自 University of California, Davis 的课程

SQL for Data Science

1486 个评分

University of California, Davis

SQL for Data Science

1486 个评分

As data collection has increased exponentially, so has the need for people skilled at using and interacting with data; to be able to think critically, and provide insights to make better decisions and optimize their businesses. This is a data scientist, “part mathematician, part computer scientist, and part trend spotter” (SAS Institute, Inc.). According to Glassdoor, being a data scientist is the best job in America; with a median base salary of $110,000 and thousands of job openings at a time. The skills necessary to be a good data scientist include being able to retrieve and work with data, and to do that you need to be well versed in SQL, the standard language for communicating with database systems. This course is designed to give you a primer in the fundamentals of SQL and working with data so that you can begin analyzing it for data science purposes. You will begin to ask the right questions and come up with good answers to deliver valuable insights for your organization. This course starts with the basics and assumes you do not have any knowledge or skills in SQL. It will build on that foundation and gradually have you write both simple and complex queries to help you select data from tables. You'll start to work with different types of data like strings and numbers and discuss methods to filter and pare down your results. You will create new tables and be able to move data into them. You will learn common operators and how to combine the data. You will use case statements and concepts like data governance and profiling. You will discuss topics on data, and practice using real-world programming assignments. You will interpret the structure, meaning, and relationships in source data and use SQL as a professional to shape your data for targeted analysis purposes. Although we do not have any specific prerequisites or software requirements to take this course, a simple text editor is recommended for the final project. So what are you waiting for? This is your first step in landing a job in the best occupation in the US and soon the world!

从本节课中

Modifying and Analyzing Data with SQL

In this module, you will be able to discuss how to modify strings by concatenating, trimming, changing the case, and using the substring function. You will be able to discuss the date and time strings specifically. You will be able to use case statements and finish this module by discussing data governance and profiling. You will also be able to apply fundamental principles when using SQL for data science. You'll be able to use tips and tricks to apply SQL in a data science context.

Module Introduction1:14

Working with Text Strings7:24

Working with Date and Time Strings5:25

Date and Time Strings Examples5:29

Case Statements7:14

Data Governance and Profiling5:43

Using SQL for Data Science, Part 15:48

Using SQL for Data Science, Part 26:14

与讲师见面

Sadie St. Lawrence
Data Scientist at VSP Global
Founder and CEO Women in Data (WID)

Welcome, we're going to shift gears a little bit in this video to

talk about perhaps a less exciting but

vitally important topic with you, data governance and profiling.

Now, data governance and profiling is not something that we're actually using SQL

for in the sense that I'm going to teach you a new function or anything.

But it's so important when it comes to writing really clean and quick queries

where you're getting this data and how you actually want to get it back.

So I want to touch on this a little bit because although this class is about SQL

for data science, it's really about the application of using SQL for data science.

An important part of that is understanding data governance and

being able to profile your data.

So after this lesson, you should be able to define data governance and profiling,

explain the importance of data governance and profiling in your data appropriately.

And discuss some methods to profile your data.

Okay, let's define data profiling first.

Profiling your data is where you're looking at either descriptive statistics

or different information on the data.

And the reason why this is important is because we've talked about how you

need to understand your data first, before you start querying it.

So profiling is a great step to begin to understand your data.

It's really simple things that you can do to start to profile your data.

Some of the things you can do is start with just understanding how

many rows are in the table.

You can also look at when was the object last updated?

Meaning when did the data get refreshed and reloaded?

Because this may change some of your results.

This may change the data that you need to limit it by.

If it's getting refreshed nightly, or in real time, do you need to set a certain

date parameter or are you okay with a consistent flow of data?

So understanding these things is essential for understanding how you might want to

pare down your data to get what you're looking for.

You can also do some column data profiling.

So for this example,

just start off with looking at what is the actual column data type?

Is it a date?

Is it a timestamp?

Is it a date stamp?

Is it a string or an integer?

This is going to change how you write your queries, for

what functions you're going to be able to use as well.

Although another thing to look at is how many distinct values are in this column?

How many rows have no values in them?

How many NULLs are there?

Is this something you need to be concerned about and deal with?

Or is there some other issue, and

some of the functions you're working with will take care of that?

And then you have your simple descriptive statistics, minimum,

maximum, average, standard deviations.

Things like that are going to be really helpful

in getting you to know your data and profiling it and understanding it.

And the reason that this is important is

because you need to be able to test your data along the way.

So as you start to write your queries,

test to make sure you're getting the results back that you expect.

If you're doing a left join from a table that has 100 rows on the left side, and

then 50 on the right side, and you only get 50 rows back.

Well, you know you did something wrong,

because it should bring back everything from the left side.

It should at least have the hundred rows.

Understanding your data can help you in testing.

And that's why using some of these simple things to look at, such us number of rows,

distinct values, minimum and maximum, can really help you get around this space.

And get you familiar wiith writing the correct queries in the long run.

In terms of governance, this is really dependent upon what

the data strategy is at the company you're working for.

Some have really strict governance policies, and some are more open and free.

In terms of a data site using SQL, it's important to understand what your read and

write capabilities are in the different environments.

Is there a sandbox you can play in and

work in to do some of the transformations and things like that to your data?

These are good questions to ask.

This is really more so

getting in contact with whoever manages this at your organization.

But understanding the governance around the data and what you're actually

able to do is important to know to keep your environments clean.

This should go without saying if you've done any programming, but

clean up after yourself as you start to write.

As you start to really explore your data, you quickly type something and

then clean something up and then try something else.

Just really keep what you're doing and keep your work really clean.

And then also understand what the promotional process is through your

environments.

If you created a model and you want to give those predictive scores and

write them back to the database, what environments do you have set up?

Is this something that you're gonnabe able to go through development, acceptance, and

production?

What is the process for that and what are you writing it to?

These are just a few little tips on how profiling your data and

understanding governance regulations around them.

But it's definitely important to look at this at your organization and

find out what these governance policies are.

So that you know why something may or may not be working.

Then you'll have the foresight to be able to write your queries and extract and

insert your data into a table in the way your organization views as best practice.

探索我们的目录

免费加入并获得个性化推荐、更新和优惠。

Coursera 致力于普及全世界最好的教育，它与全球一流大学和机构合作提供在线课程。

© 2019 Coursera Inc. 保留所有权利。

通过 App Store 下载

通过 Google Play 获取