In this Udacity Predictive Analytics Nanodegree Review, I will talk about my journey with Udacity.

*Udacity is* *offering personalised discount*.

Hello everyone, I am Mikateko. I am an **IT Risk Assurance Data Analytics Associate **at one of the 4 big accounting firms. I am a **B.com **Mathematical Statistics graduate from **Stellenbosch University **in South Africa. In this post, I will give my Udacity Predictive Analytics Nanodegree Review.

It is impossible to exaggerate the value of data in corporate management. Unfortunately, the majority of business owners lack the knowledge necessary to effectively analyze data and come to informed judgments. Tools for data analysis can be used in this situation to project future consequences and give a thorough analysis of business activities.

I am always determined about my career growth and that’s the reason why I always take valuable steps to ensure that I keep myself growing and updated.

I am still young and my mind is fresh which is the reason why I have been taking every opportunity that presented itself in my life. I have come to know about **Udacity** through my employer and initiated my journey with the **Udacity Business Analytics Nanodegree**.

Must checkout -> Udacity Business Analytics Nanodegree Review 2022

After graduating from my first Udacity Nanodegree, I was impressed with the quality and the practicality of the program which makes me decide to enroll further and study **“Udacity Predictive Analytics for Business Nanodegree”**.

There are plenty of Udacity reviews online, but not many written by students who have successfully completed the Course.

*If you buy the course through the links in this article, we can earn some affiliate commission. This helps us to keep this blog up and running for your benefit.*

Table of Contents

**Course Structure and Projects**

Let’s discuss the course structure along with the projects in this **Udacity Predictive Analytics Nanodegree Review**.

**Lesson 1: Welcome to the Program**

So, the first lesson of the Nanodegree. This was an introductory lesson. Giving you an idea of the orientation, let’s discuss the **program structure** and the structure of the projects.

**Project 1: Predicting Diamond Prices**

Predicting Diamond Prices was the **first project** which tested our understanding of the program structure and also familiarizes us with the project structure.

In this project, students need to find a **linear regression equation** in order to predict new diamond prices based on a given sample of the priced diamonds then based on this prediction give a recommendation of the bid price given the desired profit margin.

### Lesson 2: Problem Solving with Analytics

So, this lesson was more of **Linear regression** and diving deeper into its technique which is relatively easy depending on a student’s knowledge.

Further, this lesson is split into 3 sub lessons which are as follows –

- The Analytics Problem
- Selecting an analytics framework
- Linear Regression

Additionally, there is a practice project which is really essential as it provides the framework for completing the second project.

Personally, I found them a bit easy and quickly went through them.

**Project 2: Predicting Catalog Demand**

The business problem statement here was: A company that sells high-end home goods has 250 new customers and they want to see whether sending these customers a catalog will lead them to buy enough of the goods to cover the cost of sending the catalog and increasing their profit by $10 000. So the task was to predict the expected profit from sending the catalog to the 250 customers and make a recommendation to the company’s management.

This was done on Alteryx and predictive techniques which has to apply as seen on the below screenshot:

**Lesson 3: Data Wrangling**

So the third lesson was Data Wrangling.

This lesson was a combination of 7 sub lessons. These 7 sub lessons are as follows –

**Understanding Data:** The main motive of this lesson was to ensure that students understand the most common data types along with various sources of data.

**Data issues:** This lesson introduces students to common types of dirty data and how to make adjustments to the dirty data in order to prepare the dataset. You will also learn how to identify and adjust the outliers.

**Data Formatting:** In this lesson, we have learned how to summarize, cross-tabular, and reformatted data in order to prepare it for analysis.

**Data Blending:** In this sub lesson we have learned how to join and union data from different sources and formats.

**Practice Project:** In this lesson, you will get a chance to put all the skills and learnings in the previous sub lessons to a test with a not graded not submitted practice project.

**Selecting Predictor Variables:** In this lesson, we have learned the way to determine the best predictor variables that are used to predict a target variable.

Personally, these sub lessons were very interesting as they provided me a different view of the data than what I was used to. I would give these lessons an impressive score of 5/5.

**Project: Create an Analytical Dataset**

So, this was the third project, and we were needed to use all the **data cleansing techniques** we have learned in the previous section to clean the provided data and when the data meets the requirements, then build a predictive model to predict the optimum location to open a new pet store. After that, we have to put this together in a report for the pet store management.

You can see the example answers of the project below:

**Lesson 4: Classification Models**

This lesson was as interesting as all the past lessons were. However, this lesson was a bit short with only **3 sub lessons**. Those 3 sub lessons were as follows –

**Classification Problems:**In this lesson, we have learned how classification modeling differs from numeric data.**Binary Classification Models:**In this lesson, we have learned to build logistic regression and decision tree models. We have learned how to use Stepwise to automate predictor variables selection and how to score and compare the models and interpret the results.**Non-Binary Classification Models:**In this section we have learned and built to compare boosted models and forest models. We have also learnt how to score, compare and interpret results of non binary modules.

**Project: Predicting default risk**

The problem in this project was a bank received an influx of loan applications and we needed to build and apply a classification model to provide a recommendation on which loan applicants can be approved by the bank.

**Lesson 5: A/B Testing**

So, up to this point, we were focusing on predicting variables while being data-rich. In this lesson, you will learn about **A/B testing** which is all about prediction when we do not have enough data to use towards prediction models.

This lesson has **4 sub lessons** which are as follows –

- A/B Testing fundamentals
- Randomized Design Tests
- Matched Pair Design Tests
- Practice Projects

This lesson was focused on setting up the experiments and testing the hypothesis you have in mind. But this lesson was more difficult than others because I was used to **data-rich situations** and have never encountered this type of test before.

A/B Testing Guide: What is A/B Testing?

**Project: A/B Test, A New Menu Launch**

The statement of the problem in this project was like a chain of coffee shops is considering launching a new menu and I needed to design and analyze an **A/B Test** and write up a recommendation on whether the chain should introduce the new menu or not.

Examples of the project are given below:

**Lesson 6: Time Series Forecasting**

Prior to this section, the data did not need to be **time-dependent**. Time series deals with using historic data (Not like predictor variables as in the previous sections), in order to predict **future numeric values**.

What is Time Series Forecasting?

In this lesson, we have learned two types of Time Series models namely **ETS **and **ARIMA**. We have learned how to interpret their results to choose the best model for your case.

**Project:**

Actually, there was no formal project for this section, however, there was a practice project as a final project which incorporates time series modeling.

**Lesson 7: Segmentation and Clustering**

So this was the final lesson of the **Udacity Predictive Analytics for Business Nanodegree **and this final lesson was made up of 6 sub lessons which are as follows –

**Segmentation Fundamentals:**This sub lesson was aimed at differentiating between the topics such as localization, standardization, and segmentation.**Preparing Data for Clustering:**We learned how to scale data to prepare a dataset for cluster modeling and how to select variables to include based on the business context.**Variable Reduction:**We learned how to use principal components analysis (PCA) to reduce the number of variables used for cluster modeling.**Cluster Model:**We learned how to select the appropriate number of clusters and use those to build and apply a k-centroid cluster model.**Validating and Applying clusters:**Like in each and every model building, the cluster models also need to be validated. This sub-lesson was aimed at instilling model validation techniques.**Data Visualization in Tableau:**The best data scientists are those that can effectively communicate their results to an audience with no modeling knowledge. The program ended by ensuring that the students are at a minimum all-rounded predictive Analysts through instilling data visualization skills to go with the new skills learned.

**Project: Combining Predictive Techniques**

Personally, I found this project very challenging. I did not complete the project in one sitting, it took me a lot more than that. However, I manage to successfully complete the project using the mentor help functionality available within the platform.

I have referenced the project examples down below:

Check this out -> My Experience: Udacity Review 2022| Are Nanodegrees Worth $1400?

**Udacity Predictive Analytics Nanodegree Project experience**

Overall, I found the projects interesting and challenging. Initiating in a good way, as the level of the difficulty increases you will have to invest more time, concentration, and energy towards the project, which also means that you will be able to remember the **concepts** which you have learned.

My projects were usually reviewed within a day, and this project review was personalized. In cases where I did not pass my projects on the first try, the reviewer made sure that he or she clearly points out areas of improvement with classroom lessons as referrals which easily helps me refresh the concepts.

There is a specific rubric for each project which helps a lot to direct a student on the right structure to ensure they **maximize** their chance of passing the project.

**Pricing**

Personally, I don’t know the pricing of this **Udacity Predictive Analytics Nanodegree **as my Nanodegree was paid by my employer. I remember that I had to access the platform for about 4 months.

## Udacity **Predictive Analytics Nanodegree Timeline**

I remember I had to access the platform for 4 months and managed to complete 2 Udacity Nanodegree’s on time. That’s the reason why I believe the timeline is accommodative and attainable.

**Udacity Predictive Analytics Nanodegree Features**

So, there were various features of this **Predictive Analytics Nanodegree **that I liked. But still, there are some features that I liked the most and would like to tell you about them in this Udacity Predictive Analytics Nanodegree Review.

**Mentorship:**

Honestly, I did not find these features more helpful or useful than I did for the previous Nanodegree’s. Personally, I found that the concepts were too technical and as they say, they are many ways to skin a cat, different mentors had different approaches to solving the problem which can get a bit confusing for a student having a follow-up question. As a solution, I advise that the mentors have an agreed way of solving or tackling a question which I believe will solve the above issue.

**Project Reviews: **

This feature was good. I really appreciate how the reviewers are encouraging. This encouragement boosts your energy as a student even if you have not met all the requirements and still need to make the changes, you do so with an uplifted spirit.

The project reviews are very detailed and personalized.

Apart from these, there was an option of career services but I did not make use of it.

Also see – My Experience: Udacity AI(Artificial Intelligence) for Trading Review 2022

## Pros and Cons of the Udacity Predictive Analytics Nanodegree

Now, let’s discuss the pros and cons of the Predictive Analytics Nanodegree in this Udacity Predictive Analytics Nanodegree Review.

**Pros: **

Apart from the practicality of the degrees, I really like how detailed the lessons are, they are simple to follow with the simplest grammar for the students for better understanding.

I also like the idea of having a project at the end of every major concept and lesson covered as I believe is the best way to cement the concepts.

**Cons:**

I like almost everything about **Udacity**. But I would say an area of improvement would have to be the project deadlines reminders.

You need to get to the platform to be able to see all your **deadlines **which means it is easier to miss those deadlines. I would suggest that there is an automatic reminder email sent to the students maybe five days before the deadline.

**Conclusion:**

I have shared my views and experience in this **Udacity Predictive Analytics Nanodegree Review**, hope you like it. If you want to take your career to the next level, try at least one of the Udacity Nanodegrees. There is a variety of courses to take and I’m almost certain that they will carry your career to new heights.

I am a Mathematical Science graduate from Stellenbosch University and I have majored in Mathematical Statistics and Economics.