Welcome to the Statistics with R Capstone course

Loading...

En provenance du cours de Duke University

Statistics with R Capstone

105 notes

Duke University

Statistics with R Capstone

105 notes

Cours 5 sur 5 dans Specialization Statistics with R

The capstone project will be an analysis using R that answers a specific scientific/business question provided by the course team. A large and complex dataset will be provided to learners and the analysis will require the application of a variety of methods and techniques introduced in the previous courses, including exploratory data analysis through data visualization and numerical summaries, statistical inference, and modeling as well as interpretations of these results in the context of the data and the research question. The analysis will implement both frequentist and Bayesian techniques and discuss in context of the data how these two approaches are similar and different, and what these differences mean for conclusions that can be drawn from the data. A sampling of the final projects will be featured on the Duke Statistical Science department website. Note: Only learners who have passed the four previous courses in the specialization are eligible to take the Capstone.

À partir de la leçon

About the Capstone Project

Welcome to the capstone project! This week's content is an introduction to the project assignment and goals. The readings in this week will introduce the data set that you will be analyzing for your project and the specific questions you will answer using data analysis techniques we learned in the previous courses. It is important to understand what we will be doing in the course before jumping into the detailed analysis. So we encourage you to start with the first lecture to get the big picture, and then delve into the specifics of the analysis. Enjoy, and good luck! Remember, if you have questions, you can post them on the discussion forums.

Welcome to the Statistics with R Capstone course5:37

Rencontrer les enseignants

Merlise A Clyde
Professor
Department of Statistical Science
Colin Rundel
Assistant Professor of the Practice
Statistical Science
David Banks
Professor of the Practice
Statistical Science
Mine Çetinkaya-Rundel
Associate Professor of the Practice
Department of Statistical Science

Welcome to the last course in the statistics with our specialization.

Congratulations on getting this far.

Now you get to put the skills you've acquired and

the methods you've learned in the previous courses into a thorough data analysis

project, using real data to answer a concrete business problem.

>> You are a consultant hired to analyze data for a real estate investment firm.

Your assignment is to help provide insight into how the firm should invest and

to quantify and

communicate what types of real estate properties are good investment and why.

Some tips, it's generally a bad idea to buy the most expensive house on the block

and remember, the real estate agent's mantra location, location, location.

Keep in mind that the goal is to make money for your investors and hence,

investing in a property that is overvalued is rarely a good idea.

And this means that it is critical to know which properties are overvalued and

which are undervalued.

The regression techniques that you've learned in the prior courses

can help you identify these types of properties.

>> We will give you three data sets, a subset for

training, a subset for testing and a third subset for validation.

You're asked to build your model or models initially only using the training data.

Then you will test your model on the testing data.

And finally, validate using the validation data.

Could you use just all the data from the beginning?

Yes. But should you?

No, we are challenging you to keep your analysis experience realistic.

And in a realistic scenario,

you would not have access to all three of these data sets at once.

So why are we giving them all to you at once?

Well, because we want you to be able to progress through the capstone project

at your own pace, which means some of you might get to a later step in the analysis

earlier than others.

>> Over the following weeks, you'll work on a series of assignments and analyzes,

leading to the final assignment.

First, you will complete a quiz,

which will be similar to the lab quizzes you've taken in prior courses.

You will load the initial data set into R and answer a series of questions ranging

from straightforward data manipulation to more advanced modeling questions.

These questions are designed to lead you in the right direction for

your final project.

You will submit your answers to these questions in a multiple choice quiz and

receive instant feedback.

You should use this feedback to determine whether you are on the right track for

your final project or

whether you would benefit from reviewing concepts from the prior courses.

This is also a great time for discussion on the course forums.

We strongly encourage you to take advantage of the fact that you are not

alone in the course and ask questions, answer each other, and discuss findings.

>> Next, you'll complete a peer assessment where you will do exploratory data

analysis modal selection and model evaluation.

The questions on this peer assessment will be similar to the ones you worked on in

the previous quiz.

You will complete this assessment by writing a data analysis report

with R Markdown, and submit your source code and your compiled report for

evaluation by your peers, just like the final projects in the prior courses.

Your submission will be reviewed by three of your peers, and

you in turn will review the work of three of your peers using the rubrics provided.

>> Then you'll repeat this process with another quiz and then,

a final peer assessment.

For the second round, you will need to do more advanced modeling.

However, the process and

logistics will be very similar to the first round of assessments.

In your final write-up, you should provide a summary of the data and problem and

discuss objectives for your models.

You're welcome to use Frequentist or Bayesian approaches, as you see fit.

We very strongly encourage you to try both and compare and contrast your findings.

Your interpretations should especially discuss how housing prices change with

changes in covariance.

You might also want to evaluate the predictive success of your models and

provide a discussion around that.

Lastly, it is a good idea to include a summary of limitations of your models and

methods, as well as what you would do if you had more time,

based on what you have discovered so far.

>> Along the way, you might feel like you take two steps forward and one step back.

This is entirely natural.

So please do not get discouraged.

Data analysis is an iterative process.

You might have some hypotheses about the relationships

between the variables you're working with and the only

way to find out if these hypotheses are supported by your data is to explore them.

Depending on what you discover during the exploration,

you might decide not to include it in your final project.

This does not mean you wasted time.

In fact, if every single visualization you made and every model you

fit makes it into your final project, you're probably making a mistake.

You want to try a variety of approaches and then pick and

choose between them to craft your final project.

That being said, you probably do not want to

constantly feel like you're going down the wrong path or

having to backtrack too far in your work due to a careless mistake.

In order to avoid this, we strongly recommend that you take cues from

the answers to the quiz questions and peer assessment feedback along the way.

>> Good luck with your analysis, and don't hesitate to ask questions and

share your findings in the course forum.

Keep in mind, there is no one right answer and your justifications and conclusions

are just as important as the methods and techniques you use to arrive at them.

Explorer notre catalogue

Rejoignez-nous gratuitement et obtenez des recommendations, des mises à jour et des offres personnalisées.

Coursera propose un accès universel à la meilleure formation au monde, en partenariat avec des universités et des organisations du plus haut niveau, pour proposer des cours en ligne.

© 2019 Coursera Inc. Tous droits réservés.

Télécharger dans l'App Store

Disponible sur Google Play