Selecting a right course is one of the top most dilemma that you come across when you are starting out on your data science journey. As the first step of your journey, you need a guided course which starts from basics and take you through advanced algorithms of machine learning. Therefore, it is very important to select the right course. The universal truth is that there are very high chances of giving up the subject in the beginning only if you find the subject incomprehensible. That’s why a right starting approach is very important. Also, there might be chances that you are not very sure about becoming a data scientist and you think that you will just go through a course and half way through you will decide if data science interests/suits you or not depending upon if you are able to feel the connect with it and are able to understand the concepts. And, just in case you opt for a course which does not explain the concepts well and leaves them incomprehensible for you then there are very high chances that you would give up data science thinking that its too tough or its not for you. Who is at the losing side then? You & only you. Because you missed out on such an awesome opportunity of having a great career in data science. That’s why I give so much importance to selecting the right course, specially when you are just starting out.
I was at the same stage around 5-6 years back where you are right now. I was a sales guy and was very much fascinated by data science field. I had just heard about the application of data science. Concept of solving business problems and taking big decisions by drawing inferences from data by applying machine learning algorithms sounded very enthralling to me. I was so fascinated by data science & machine learning that I just wanted to learn them as soon as possible. But, I did not know from where I should start. I wasted months by joining and leaving tens of courses mid way after realizing that I am not able to understand much. Input that I was giving was much more than the results I was getting back. After wasting so much time, I realized that it was very tough for me to keep my learning enthusiasm high. I would give up if I don’t find a good course and make a breakthrough soon. I kept on searching, joining and learning these courses. Fortunately, I came across some good courses also with time which equipped me with data science knowledge. In this entire process of months of searching for a good course, I figured out 5 most important and basic parameters of selecting the right course. These 5 most crucial parameters are as below:
- Course structure
- Teaching methodology & concepts explanation
- Real-world data
When you are just beginning to learn data science, you need to see the width of coverage than depth in these courses. Why? Because you are not going to be a subject matter expert of one topic in data science in the starting of your career. You don’t even know of you if you like text analytics more or number analytics more. You develop liking/disliking for a particular area only after working on that for some time. That’s why you first need to get exposure of most of the areas in data science. You must know that what all types of problems can be solved using data science and what all algorithms are used to solve them. You don’t have to restrict yourself to just one type of concept or business problems. That’s what is expected from a fresher or no data science experience candidates in an interview. They should have little knowledge of most of the concepts/algorithms so that company can train them more for their specific work. But if you have no knowledge of the concept at all which company is looking for then it would be very difficult for company to take you. So, rather than going deeper into particular area, try to cover width of data science. At the later stage of career, you can opt for a particular field in data science and become a subject matter expert of it. In sum, go for a generic course than a specific course in the beginning.
To cover a good width of data science, a course must include or touch base upon the following:
- Introduction to Programming Language (like Python, R)
- Fundamentals of Statistics
- Descriptive Statistics
- Inferential Statistics
- Hypothesis Testing
- Machine Learning
- Introduction to Machine Learning Concepts like
- Feature Selection
- Parameter Metrics
- Regression Algorithms
- Linear Regression
- Ridge Regression
- Lasso Regression
- Classification Algorithms
- Logistic Regression
- Decision Tree
- Random Forest
- Clustering Algorithms
- Centroid Based
- Connectivity Based
- Density Based
- Text Analysis
- Natural Language Tool Kit(NLTK)
- Naive Bayes Theorem
- Bag of Word
- Sentiment Analysis
- Introduction to Machine Learning Concepts like
Just go through the content of the course you are considering to take and if you find all above topics in the course curriculum that means course covers important and essential concepts and satisfies this criterion.
Teaching methodology explanation of concepts & personal touch
Data science or any tech concept is learned in a process. Any new subject has to start with simple basics, take you to intermediate level and finally advanced level. All this happens without you noticing it consciously. Your level keeps on increasing automatically during learning the course. But what would happen if you are directly exposed to advanced concepts? The course will make a good impression on you that it’s teaching something very advanced, but will you be able to learn anything? Understand this by drawing an analogy with learning a guitar. Just imagine that you are just beginning to learn a guitar, and I give you chords for a beautiful but advanced level song, will you be able to play that? You may get impressed with my skills, but you will not be able to gain that skill. Rather, you will be so discouraged while trying to learn that toughest song chords that you will give up on learning a guitar and make a mental barrier that you would never be able to learn a guitar. On the other hand, there are students who start with simple 2-3 chords song, learn and practice them well. They will keep on raising their level gradually by playing tougher songs and will be able to tough songs with ease eventually.
As I told in my other blog also that there are some of my friends who joined courses which talked about very advanced concepts which looked very impressive from outside, but when they actually started learning them, they could not understand them and got demotivated and left data science forever.
Explanation of Concepts
There is a very famous saying :
You do not really understand something unless you can explain it to your grandmother.
I love this saying so much. This does not apply to students, rather it applies to courses teaching data science. These courses should explain the concepts in such simple and laymen terms that even a novice does not feel alien. I don’t like over use of technical jargon which many courses do to show authority over the subjects. But, unless and until a course connects with masses and beginners, knowledge will not flow to them. It is very important to make a connection with students, and that happens only when students feel comfort while learning. They should not be put in a loop where to understand a concept they come across two other unknown concepts and first they have to understand them, and two learn these concepts they have to understand few more concepts. This is the most frustrating thing I have come across while learning technology. If you try to understand one technical line, there will be a couple of terms which you will not understand and then you have to understand these terms and it goes on for long before you understand the original concept. It’s actually very exhaustive process and discourages students to move ahead with the course.
That’s why I emphasize so much on how a course explains the concepts. You can get this feeling by going through free preview chapters or mini course offered. Just go through them and see if you are able to make a connect with the course. Do you think that you don’t need to read one line again and again to understand it, which is the case with most of the courses offered? My simple way of opting for a course was if I find any of two underlying things in the course when a new concept is introduced.
1. Analogy :
This is the most powerful thing in explaining any new concepts to a beginner who does not know about the new concept/field. Students come from various backgrounds and may not relate to the explanation if it’s purely in technical terms and uses jargon. Explaining the concept by drawing an analogy with a daily life thing which a student finds very relatable, makes the difference. This is the single most important factor which makes a course stand out in my eyes.
2. Examples :
This is another very critical factor which differentiates a good course from the crowd. Every concept must be followed by an example. If a technical concept is explained, it must be followed by a small code snippet which is simple & easy, and does supplement the explanation well. It is like when you complete some task, you give it a finishing touch and task is considered complete. Similarly, an example should work as finish touch and after that you should be able to understand the concept well. If you don’t feel confident about the concept even after going through the example, that means there is something wrong with the course. Please not that this won’t be true and some concepts will require you to put more efforts or read them again, but if it’s happening very frequently that means course is not worth pursuing.
Now, we come to the most important criterion which is very close to my heart. I would any day prefer a course which makes me feel at home. The course should give me a feeling that somebody is explaining the concept form heart and really wants me to learn. I should feel connected to the course. I should feel like going to course again and again to learn the next chapter. It should also create enough excitement in me to finish the course and learn as much as possible.
This can’t be measured using a quantitative unit, rather it has to be felt. There is no quantitative or tangible parameter which I can tell you to judge the course. You have to experience it by yourself by going through the free material given by course. If you feel the personal touch in the course, then don’t think twice, just go for the course. This may vary from student to student. One course that gives a personal feeling to a student, may not do the same with other student.
Your learning is incomplete if you can not apply what you learnt to solve a problem. Exercises are integral part of any course. I came across many video courses which do well in covering the topics and explaining the concepts but they fall short when it comes to this criterion. These courses will make you feel confident as long as course is open in front of your eyes. As soon as you close the course, you start feeling under-confident. You open the course again and go through the material; you start feeling confident again Haven’t all of us experienced it? It happens because we don’t practice on new problems. Believe me, once you solve a problem using newly learned concept, your confidence will skyrocket. Exercises inherently do three main things:
- Check how much you understood the concept – You don’t know how much you understood the concept well till you are able to solve a problem using the concept independently. Solve a problem and you are done with the concept. An exercise is like a certificate that you get for successful completion of a concept.
- Make you more confident about your understanding of the concept – just learning a concept theoretically is not enough. A concept learnt theoretically does not stay with you for long unless you practice it. After sometime, a theoretically learned tech concept is as good as not learnt. If you want to feel confident about a concept and want it to stay with you for a very long time, do exercise on that concept.
- Increases Recall value – Have you heard about Brand Recall? A company’s brand value is directly associated to its presence in the memory of consumers. If you ask them a question indirectly related to brand, can they recall the company name or does the company name come into their mind naturally? That’s brand recall. When you solve a question using a learnt concept, its brand recall value increase in your mind. If you come across a new problem in future which can be solved using the learnt concept then this concept will automatically come into your mind. This will happen when you actually solve and exercise using the learnt concept. It puts a longlasting impression on your mind about the concept.
Therefor, it’s very crucial that
Why are you learning data science? Why have you joined a data science course? Because you can apply them in your company or solve a real-world problem. Right? So, it’s very much required that you work on real-world data in the examples or exercises given in the course. Make sure that the course you are going to join, provides you real-world data to practice. If, in future, you are going to work on clustering(segmenting) the customers into some categories so that you can server them well by customizing your services according to their needs, then you should work on real-world data now in your course so that there is no surprise factor when you actually work on the data in future at your workplace. There is a very famous saying:
the more you sweat in training, the less you bleed in battleground
I have interviewed candidates who have no working experience in industry and have just completed a data science course. Sometimes, these candidates have much better knowledge than people working in the industry. The main reason is that these guys have actually worked on real-world data provided by the courses. They don’t feel any discomfort after joining the company, and they hit the ground running. They deliver the results as soon as they join the project in company.
This is another key criterion which makes it to the list. The course which you have joined should not leave you in lull. There are generally two two types of issues where a student needs support:
- Technical support
- Learning support
It mainly includes support related to access of learning resources – text, video or availability of course website. You may face issues like you are not able to watch videos, site is down, page is not loading, etc. These are generally less relevant in today’s time because technology has improved a lot in recent time and apart from rare glitches, courses are generally up and running all the time, with smooth access to learning resources without any major issues.
I would like to emphasize more on this. Sometimes, you are not clear about concepts explained in the course and you want a better explanation. In such scenario, what should you do? Does course provide any help apart from the course material. How fast you get the response?
Another area where you might need help from course provider is exercises. If you are not able to solve the exercises somehow, what support course provides you. How fast it can provide you with solutions so that you don’t get stuck on one topic and can quickly move on after getting the solutions?
Both technical and leaning supports are non-negotiable. You should try to find answers of above questions before joining any course.
I will conclude the article now. Just don’t join any course. Do a due diligence on above mentioned criteria about the course before joining it. Selecting the right course to start your data science journey is a very important step that many people ignore. What kind of data scientist is going to come out of the course depends on what type course it is. Keep the following in mind :
Good data science courses produce good data scientists