Thayer School of Engineering at Dartmouth | Bootcamps

It took a while for me to pull the trigger on enrolling. What sealed the deal was the 50% tuition rate offered, which cut the price from what felt like an outrageous $12.5k to a much more reasonable $6k. I was out of work (at the start of the course) due to the pandemic so it felt like a good time to jump in.

The course begins with an entry math quiz, which was reasonably difficult. It included things like linear algebra, high-level calculus, and advanced statistics. I had never taken linear algebra before but was able to learn quite a bit with Khan Academy. You don't need a perfect score on this entry quiz, but you do need to score around 80% to my recollection. 

The course begins with a month-long Python coding bootcamp. This portion is technically ungraded and you can totally skip it if you have a good background in Python. All the learning occurs on the DataCamp platform while homeworks are completed in Jupyter notebooks. There is a final Jupyter project for this Python coding section which IS graded. 

The rest of the course is split into 10 modules. Each module consists of two weeks. The first week consists of recorded lectures, some required textbook reading, a one-hour live office hour, forum discussions, and a coding assignment. The second week typically consists of math-based video lectures, a math quiz or two, and a live coding assignment done together as a class and then finished/submitted on your own. The modules are as follows: Practical Applications of Analytics, Data Structures and Plotting, Introduction to Statistics and Probabilities, Linear OLS Models, Linear Regressions Interactions and Transformations, Logistic Regression and Applying GLM, Data Visualization Strategies, Experimental Design/Causal Research/Targeting Analysis, Machine Learning and the final module is the final project. I found all the modules to be very valuable and a good mix of new challenges/repetition. 

The final project basically consists of identifying a potential business problem, curating an applicable dataset, and applying machine learning techniques to attempt to solve the problem. You will use Python and relevant ML libraries (mainly sklearn, but others can be used). The trick to the final project is finding something relatively novel that is also doable. There was a broad range of success/failures among my cohorts' final projects. Some people picked things that were too simple and were basically considered mainstay ML problems (ex. identifying mushrooms/flowers based on measurements) while others picked ML problems that were way too big in scope or ill-defined. 

The course is also jam-packed with career-oriented learning opportunities. There are multiple sessions with career coaches and assigned mentors, both group and one-on-ones. Be sure to take full-advantage of these, especially if you are planning on breaking into a new data science role. 

Expectations:
Expect to spend at least 13 hours a week on this course. I'm definitely not a slow-learner and I felt 13 hours was my average. Expect to spend even more time if you have any foundational knowledge deficits such as calculus or statistics. 
Do not expect to become a pro after just this course. It will give you all the necessary skills and knowledge to further your data science journey, but there is much more to do after. I cannot speak to whether it provides everything needed to become competitive for entry-level data science jobs, but I imagine it does, especially at the larger firms like Amazon, AirBnB, etc. 

The things I didn't like: 
I wish we had more access to actual Dartmouth faculty members. The only actual Dartmouth faculty member you will see will be in the pre-recorded lectures. All live interactions occur with an Emeritus employee. He does a fantastic job, but you can tell he is teaching multiple courses at once. Don't expect office hours to last more than the allotted hour. 

Secondly, I was a little disappointed in the range of knowledge/skills among course participants. It felt like far too much time was dedicated to subjects/knowledge that was included in the entry quiz. What's the point of the entry quiz if we're going to end up going over most of that stuff anyway? It just kind of felt like wasted time and I wasn't getting my full moneys worth reviewing stuff that I felt was foundational. 

Third, I wish we had spent a lot more time on machine learning as opposed to the more basic stuff. Again, basic statistics were tested for in the entry quiz. Sure it was nice to learn how to code linear regressions in Python, but the meat and potatoes should have been ML. After all, you really only spend two weeks on the ML stuff and it felt rushed since there is so much to learn in the ML world. 

Similarly, I wish we had done the ML stuff earlier in the course or at least touched on it more frequently throughout the course. Ultimately, your entire final project is based on ML. But it's a little annoying to be expected to come up with a final project when you don't even really understand the capabilities of ML, to begin with. I know I felt a little lost and in the dark.  

Overall, I highly recommend this course to anyone interested in data science. 

Why I Pursued Data Science
About a year ago, when the COVID-19 pandemic was still in its relatively early days and there was great uncertainty about the near-term future of the civil engineering industry, I decided to enroll in the Professional Certificate in Applied Data Science (PCADS) program at the Thayer School of Engineering at Dartmouth to open a pathway for a career pivot into Data Science. I've always enjoyed solving problems analytically using math or simulated methods, and I knew that I had thoroughly enjoyed MATLAB in college. Data Science is an emerging and exciting career path that is not confined to any one market sector or industry. 

Typical Options for Entry into Data Science
For those looking to pivot into Data Science, you may notice that you can go one of several typical routes:

1. Enroll in a Masters program.
2. Enroll in a bootcamp.
3. Teach yourself skills over a long period of time via mass open online courses (MOOCs) offered by sites such as Coursera or DataCamp.

The problem with these choices is that none of them provide an expeditious, cost-effective path for someone looking to transition careers while still working or those with other life obligations. Masters programs are very expensive (often $60k+) and can take 1-2 years on average. Bootcamps are in that mid-range ($6-20k) in terms of cost and they can be completed quickly (~3 months), but they seem to only cover the superficial topics in Data Science such as the 'how to' code in Python or navigate around certain open source machine learning packages such as Scikit-Learn (sklearn).  This is valuable, but it seems to me that a Data Scientist ought to know more than just how to use the current industry software. As for the MOOCs... this could be a good option for some and is most definitely the least expensive, but it would require a tremendous amount of self-motivation and they are known for taking years to amass a number of course certificates that don't necessarily translate into career results.

I needed a fourth option...

"The Fourth Option" :   PCADS
I chose the PCADS program because it was offered at a cost comparable to that of a Bootcamp, but it included all the highlights of a typical bootcamp and the core mathematical principles that one would learn in a Data Science masters program. The Dartmouth PCADS program also came with a few additional perks, including access to a career mentor, guided career learning objectives as graded assignments (motivation to work on career development early and often), technical presentation guidance and samples, and a 1 year unlimited subscription to DataCamp.  PCADS is a good hybrid between a full Masters degree and a bootcamp.  I often refer to it as a "mini masters" -- it really is!

The Program
To enter the course, one must complete a graded test on the subject of calculus, linear algebra, statistics, and probability.  My undergraduate degree was in Civil Engineering with a Minor in Business Administration, so I have had plenty of formal education on these subjects.  That said, it had been over 6 years since I had thought about vector calculus, matrix determinants, or kurtosis, to name a few.  I studied for an entire month before taking the entry quiz for admission into the course and I strongly recommend serious candidates plan for a month of self-study prior to taking the entry quiz and starting with your cohort.

The PCADS progam is a a six (6) month program.  Once in the program, four (4) weeks are provided to complete eight (8) courses on DataCamp, and then a final graded bootcamp assignment on Vocareum.   Technically speaking, only the graded Vocareum assignment is required.  However, if you actually seek to complete all provided materials, you will need to allocate much more than the course's purported "15-20" hours per week.  I think I spent about 30-40 hours per week during the bootcamp phase of the course.  It was tremendously difficult to juggle while working full time, but I have no regrets.  I highly recommend taking full advantage of every single course they recommend as prerequisite to the rest of the program.

The rest of the program is divided into 10 modules.  Modules 1-9 go over many topics such as multivariate linear regression, linear transformations and interactions, experimental design (A/B testing), data visualization, and machine learning, among others.  Each module is divided into two weeks that follow a typical format:

• Week 1 of each module introduces the mathematical concepts and includes some required readings.  It also includes a graded Python assignment where you "get your hands dirty" implementing code that utilizes the learned concepts. 
• In Week 2, you will take a test on the mathematical concepts and also work on some additional coding via a live demonstration with your course instructor (also graded).

In general, the modules provide a good mix of mediums that help keep it interesting:

• Videos
• Readings
• Interactive forums/posts
• Assignments
• Mini quizzes
• Grades quizzes
• Live webinars / office hours
• Career assignments
• Career webinars
• Presentation practice webinars

All the above are interspersed throughout the course.

The final module (Module 10), is where students complete a final project.  The final project includes a Jupyter notebook submission and a 10-15-minute final presentation recording.  Students are given great leeway as to what topic they can use for the final project.

Advice to Future Students
I highly recommend the PCADS program to those who are already technical and looking to pivot into a career in Data Science.  The course is what you make it.  You'll definitely need to spend at least the higher end of the time estimate (15 hours) per week just to pass.  But if you are serious about making a career pivot and really want to feel like you're learning each topic in depth, and if you spend the time working on the career assignments, you'll need to spend closer to 20-30 hours per week on average.  I recommend taking meticulous notes throughout the entire course and bookmarking every single link the program provides to open source articles or free books.  SAVE ALL FILES/NOTEBOOKS offered by the program--you paid for access to these so don't lose them once the course is over.  Backup all your work to the cloud!  Also, ask lots of questions during your office hours.  Again, you paid for access to a live instructor so you may as well learn as much as possible!   

Feedback to Course Instructors  
First off, thank you for putting together a course that, in my opinion, is truly a unique experience not offered by bootcamps or Masters programs!  My feedback is that the course advertised 15-20 hours per week as the time commitment, but I think a diligent student needs to put in substantially more time.  I think the four (4) week bootcamp portion should be extended to six (6) weeks and I could say the same for the final project, which really could be as much as eight (8) weeks.  For students (like me) who had to work full time throughout the course, it was very high pressure to complete the final project in such little time.  Other than the time estimate being a bit off, access to the Dartmouth career services (or alumni network) would be a great benefit.

There were some hiccups with the program support taking a while to respond (+/- 1 day) and assignments not being posted on time (reducing time for students to do the work), but things went well for the most part.

In summary, I highly recommend the Dartmouth Professional Certificate In Applied Data Science program!  Knowing how much work went into graduating the program, I will have a tremendous mutual respect for other future graduates!