First steps in Data Science: How to gain practical experience
One of the key issues faced by beginners in Data-Science is gaining practical experiences.
If you are a motivated learner who has decided to put their feet into the field of Data Science, but you lack practical knowledge on getting your hands dirty with real-world data, then this article is just for you. Here, I will talk about how to begin your journey working with Data and to land in a technical interview.
So, let’s move straight into those first steps:
1. Target An Industry
One frequent mistake of beginners is to start looking for any sort of data to put their newly learned skills into. This is fine to an extent, especially when you are still learning to apply algorithms and techniques. But knowing only traditional algorithms is not enough in this field. You need experience with industry-targeted data.
Let me explain with an example,
Say, you wish to work in the automotive sector, and previously you had taken courses on Data Analysis and Machine Learning. You have worked with “Bundesliga data” of the same year and were able to find key insights and predictions. You have also worked on the infamous “Housing Price Prediction” data applying all your Machine Learning knowledge.
In all, you think you are a good package for this automotive company — because you know Data analysis and Machine learning techniques both with experience of working with data. With this confidence you apply for the Data Engineer position — only to get the rejection. Why?
Because there was another person with the same technical skill set as yours but with experience in data in line with the automotive industry — like, “risk analysis on autonomous vehicles”. They must have also had crucial skills like working on Business Intelligence tools like PowerBI or knowledge on MLOps. This makes the recruiting company and their client spend less amount of time on knowledge transfer and in explaining data as the functioning of the industry and its jargon are already known by the candidate.
That being said, you don’t need to worry too much if you are unclear about choosing an industry, just pick one and explore until you are no more curious to get the answers (“After all, learning begins with curiosity, right?”)
In short: Gain practical experience on that specific industry dataset, you wish to work for in the future.
2. Find persisting problems
“If I were given one hour to save the planet, I would spend 59 minutes defining the problem and one minute resolving it”
— Albert Einstein
At this point, you might be wondering that experience on relevant industry data is “The only Crucial part”. Well, that is not completely true either (what is the absolute truth anyway? ;) ).
You also need to spend time understanding how that industry works. This will help you to identify the existing problems of that industry. Without a problem to solve, there won’t be any work for you!
By knowing the influencing elements of an organization and the difficulties faced you will firstly have more motivation and knowledge to attack the problem and secondly you will be definitely better off with your analytical aptitude. Lastly, if you identify a way in which you can contribute to an organization of your target industry you will become a useful resource.
In short: Understand the working, find problems, make space for yourself.
3. Do some Kaggling
“Kaggle” should not be an alien word if you are getting into the world of Data Sciences. It is the temple of learning for beginners and an experimenting playground for intermediates. If you follow the first two points mentioned above, the next step would be to find data in order to get a hands-on experience. And Kaggle has an abundance of datasets for each industry, you could simply search the domain of your interest like — healthcare, education, fintech, travel, Online Shopping, etc. Moreover, every dataset that is present on Kaggle has a description of how and from where the data was gathered, this proves helpful while drawing conclusions and making reports. Do not worry if you get stuck somewhere in your analytical journey, as the Kaggle community, and their discussion boards are always at your rescue. In the worst case, submissions by previous participants are available and can be helpful for finding good starting points, a common difficulty among beginners.
To start with download a dataset from your domain of interest and the one which drives you the most (“cause on days when you will have 1000 reasons to give up, your motivation will work as fuel”).
So, plunge into the data and create your own approach for
1. Data cleaning
2. Forming relevant questions to progress through the analysis
3. Feature Engineering
4. Visualization of your findings
5. Reporting
6. Model building and Predictions (for ML tasks)
In short: Download a dataset and get your hands dirty, don’t forget to document your work on git-hub and Kaggle profile.
4. Become competitive
Competitions/hackathons are a great way to know where you stand in terms of learning. But it is a strict no-no for absolute beginners as these competitions look intimidating without any basic grasp of the subject and required coding skills.
For learners who have some prior experience and have done at least 2–3 self-learning projects, can start their hackathon journey from beginner-level competitions. The accolades achieved in such competitions are greatly acknowledged by recruiters and the Data science community.
Data science hackathons are usually of the type, where a real-world industry problem is put forth, a dataset is provided and within a short span of time participants are required to come up with a solution, one with the highest accuracy and relevance wins. In the end, every team is required to pitch their idea and their problem-solving approach through a presentation.
As an example — you might be given a LinkedIn marketing campaign dataset of the company and asked to find out key insights — as to what worked and where the company must invest next and at what time. (Pretty real and business-critical, isn’t it?)
Some of the popular platforms for these hackathons are — Kaggle, DrivenData, DataCrunch, HackerEarth, Topcoder to name a few.
These competitions make you develop fast problem-solving skills in a cross-functional team and guide you to discover your own limitations. Not to forget the opportunity to network with the companies who sponsor these competitions as an added advantage of participation. It is also an excellent way of gaining a practical understanding of day-to-day Data scientist’s tasks for people with zero job experience in this field.
In short: Participate in Hackathons to develop your personal hacks and lend your help by answering questions on Stack Overflow and other such platforms.
5. Get ready for interviews
If you have followed all the four points mentioned above and are still hooked into this field of Computer science, you probably want to make money out of these skills, and why not!? But before you get to stand Infront of the interview room, you must set up your portfolio and CV. A Kaggle profile and GitHub repository filled with your documented projects help your profile to stand out. Add your worth mentioning hackathon experiences in your CV too.
For Interview preparation, you could use websites like Leetcode and HackerEarth. These websites have company-specific questions as well as are based on varied data science topics. Besides these statistical concepts, probability theory, fundamental modeling concepts such as regression, classification, clustering are also important and must be prepared before heading for the interview. For refreshing your statistical concepts, you can check out Brilliant.org and MIT Open courseware which are available for free and are recommended by FAANG recruiters.
In short: Prepare data science theory too, for the interviews.
These are the few points which I think, if you follow in the sequential order, you might gain a lot of necessary practical experience to form your career in the field of Data sciences. No guarantees of you excelling in the interview, but I firmly believe in what Sir Isaac Newton once said,
“My powers are ordinary. Only my application brings me success”
— Sir Isaac Newton
So, keep applying your knowledge, you never know when a wisdomous apple will strike your head 😉
