Udacity’s Data Scientist Nanodegree Review
As I received a lot of messages from my LinkedIn contacts about my thoughts on the nanodegree, I thought it would be be best to write a post about the topics, timelines and project portfolio that you will end up with and of course a final verdict if it was worth it or not.
What made me enroll ?
I took a machine learning class during my Masters and worked on some machine learning challenges at my work and internships. Looking at the growth in this field and my increasing interest, I decided it’s time to learn more. Although, there is a whole lot of material that can be self-taught and kaggle projects that can be self-implemented. I thought it would be best to take a more organized course and have someone review my data science projects while starting out.
After all this thinking, I enrolled.
The Nano degree is broken into two terms:-
Term 1:- Machine Learning for Data Science
Term 2:-Applied Data Science
Timelines
Term 1
This lasts for about three months. You will need to work about 20–25 hours per week(if you make notes).
Term 2
This lasts for about five months. You will need about 25–30 hours per week(if you make notes).
The hours can be slightly less or more depending on your expertise, learning speed and note making technique. Also, you will get an automatic one month grace extension if you fail to complete the course in the above mentioned time frame.
Cost
I was lucky to get a deal while enrolling. But, right now it costs $799 and $999 for Term 1 and Term 2 respectively. You can avoid this by getting a scholarship. I do not know how this works, but it’s listed on their website.
Course Contents and Projects
Term 1:- Machine Learning for Data Science
Supervised learning
They have covered Linear Regression, Perceptron Algorithm, Decision Trees, Naive Bayes, Ensemble Methods, Model Evaluation Metrics and Training and Tuning Classifiers.
This course ended with a classification project in which the goal was to predict whether a person would donate to charityML or not. You can find this in more detail here.
Deep learning
This includes Introduction to Neural Networks, Implementing Gradient Descent, Training Neural Networks, Keras and PyTorch.
This ended up with an Image Classifier project.
Unsupervised Learning
This includes topics such as Clustering, Hierarchical and Density Based Clustering, Gaussian Mixture Models and Cluster Validation, Dimensionality Reduction and PCA and Random Projection and ICA.
This ended with a project titled Identify Customer Segments . Here, the data set was provided by Bertselmann.
Term2:- Applied Data Science
Introduction to Data Science
This includes details about CRISP-DM, communicating to Stakeholders.
The project was to write a Data Science blogpost leveraging the CRISP-DM process.
Software Engineering
This includes software engineering practices and how they apply in data science, writing clean and modular code, code efficiency, refactoring, documentation and version control, testing code, logging and conducting code reviews, object oriented programming and web development.
This includes an optional project to deploy a data dashboard.
Data Engineering
This module includes thorough explanation for ETL pipelines, NLP pipelines and machine learning pipelines.
This modules concluded with a project titled Disaster Response pipeline.
Experimental Designs and Recommendations
This includes concepts in Experimental Design, Statistical Considerations in Testing and A/B Testing Case Study. Also includes recommendation systems and matrix factorization for recommendations.
Data Science Capstone
In this capstone project you will solve a real word problem using all that you have learnt throughout the program.
Here is my capstone project.
Final Verdict
The course is definitely worth it as it is highly structured and has covered all the important topics spanning machine learning and data science. The portfolio projects will also help hitting the keywords during job search. It is highly beneficial in case you do not have relevant work experience. In case the course fee is a barrier, a scholarship is always an option.
It definitely seems as a good option for career transition.
