From the Feynmann Technique to the Leitner System, this guide has a study technique to help everyone learn data science.
Data science remains one of the most sought after, yet most difficult to enter and master fields even into the second decade of the 2000s.
What you hear about college dropouts learning to code in three months and then making $100,000 as an entry-level salary at a FAANG company does not apply to jobs in data science.
Instead, anyone aspiring to become a data scientist seems to have been born for the job, blessed with coding, analytical, and mathematical skills that seem to produce beautiful visualizations out of thin air that perfectly describe an organization’s business problems and successes without having to as much bat an eye.
However, what does this mean for the rest of us mere mortals who perhaps came onto data science later in life or those who are looking for a career change?
Luckily, many study techniques can be used to anyone’s advantage to help them master the toughest topics in data science that will help land that first job in the field. From the Feynmann Technique to the SQR3 Method, there is a study technique in here for everyone who wants to expand their knowledge and master everything they need to break into this enticing field.
Spaced repetition is the process of “reviewing information at gradually increasing intervals”.
Take for example, that someone tells you how to calculate the mean value of a statistical sample. This information is great to know, but if you’re not using it regularly, you will likely forget how to calculate the mean. However, if you were to periodically review the concept of calculating the mean value, you would be more likely to retain the information and be able to use it more efficiently in the future.
Essentially, the brain contains synapses which are junctions between neurons. Simply put, synapses are used to conduct nerve impulses between neurons which helps the brain transmit information. Synapses that are rarely used will become weaker and may even be “pruned” by the brain. This allows more regularly used synapses to become stronger (Kalat, J. W. (2022). Introduction to psychology. Boston, MA: Cengage Learning).
Therefore, spaced repetition works by strengthening the synapses in your brain for retrieving and using specific information.
Spaced repetition can be accomplished by regularly reviewing material and retrieving that information from your brain. Creating a schedule can be an efficient way of ensuring that you are regularly reviewing the material. Alternatively, the digital flashcard tool Anki is a great all-in-one spaced repetition tool that helps you set a schedule and ensures that you are given the right amount of practice every day.
What this technique is good for:
Increasing the retention of vocabulary, methods, equations, and concepts.
How you can use it to study data science:
Spaced repetition is a great way to practice using the algorithms that you regularly use in data science. While not all data science projects use the same algorithms, it can be a good idea when first learning to practice using all of them. Additionally, it can be beneficial to practice them over time to ensure that you don’t lose the ability to quickly and efficiently implement one in your next project. Furthermore, spaced repetition of algorithms in different scenarios will help you understand how they apply in different situations and the range of results you may get depending on the algorithm you choose to implement.
Active recall is a study technique that involves “retrieving information from your brain”.
Where most people would normally consider studying to be an act of putting information into the brain (sometimes referred to as a passive study method), active recall works by forcing the brain to remember and retrieve information actively.
According to a multi-institution research study on effective learning techniques conducted in 2013, active recall was ranked as a highly successful study technique that has been shown to boost student performance across educational contests.
Active recall can be implemented through the creation of practice questions, practice tests, or flashcards that you must attempt to answer by recalling what you know. If a question is answered incorrectly, go back to your notes to understand what your answer was missing. Then, go back through the questions once again and attempt to improve your previous score.
What this technique is good for:
Solidifying your memory of topics, as well as improving the ability to quickly and confidently recall information to answer questions.
How you can use it to study data science:
Active recall is a great study tool to use when you are first getting started with the basics of data science, such as the fundamentals of programming using Python or R, learning the basic mathematics involved in data analysis, and learning which type of visualization works best for which scenario. By developing a list of questions for yourself as you move through each topic, you can quiz yourself later to see if you really understand how to use programming, mathematics, or data visualization in each situation. Not only that, but by developing your own active recall test questions, you will learn to use the information you know in creative ways and you will begin to anticipate how you would solve a problem.
Note: I have an extensive article detailing how to use the Feynman Technique that can be read in its entirety here: How to Use the Feynman Technique to Become an Expert in the Most Complicated Concepts in Data Science.
The Feynman Technique was developed by Nobel-Prize-winning physicist Richard Feynman, a pioneer in the field of quantum computing and nanotechnology, who was known as the “Great Explainer” for the great lectures he delivered at Cornell and Caltech.
The Feynman Technique is a four-step process for understanding any topic that works by developing true comprehension of a topic through active learning.
The Feynman Technique can be broken into four steps that are completed in sequence for a given topic:
- Choose a concept to learn: Write down everything you already know about a topic using your own words and brainpower. Make a note of any point where you must use a long complicated word to explain something and determine whether or not you understand what the word means. If you can explain what the word means in simple terms, great! If not, study the word and its usage to learn more deeply what it means and then try to figure out a way to describe the word using simple language.
- Teach it to yourself or someone else: Teaching a concept to someone else truly determines whether or not you understand the concept yourself. It’s very hard to trick yourself into believing that you are an expert on a topic when you can’t teach it coherently to someone else! The key is to teach the concept in such a way that a 10-year-old could understand it.
- Identify your knowledge gaps and return to the source material: After attempting to teach the concept to someone else, you will be able to identify gaps in your knowledge based on the ease with which you were able to explain the topic, and whether or not you were able to answer any follow-up questions. Once you’ve identified and filled any gaps, it’s time to go back and teach the concept to someone again. This step is part of an iterative cycle with Step 2. This cycle should be completed as many times as is necessary until you know the topic forwards and backward.
- Simplify your explanation: The final step in the process involves re-working your explanation so that it excludes jargon and makes use of simple analogies that are easy to recall and explain in the future. This step forces you to simplify the concept until it can be understood fully by anyone.
What this technique is good for:
Understanding the link between theoretical, technical, and mathematical concepts.
How you can use it to study data science:
The Feynmann Technique is a great way to deeply learn the toughest concepts of data science because it forces you to understand them to the point where you could explain them to anyone. For example, unsupervised learning or descriptive models are concepts of machine learning that can be complicated to understand because there is no exact target. However, learning how to explain how unsupervised learning works and understanding in what scenarios descriptive models can be beneficial to a basic level can help you better understand them in a simple way. Then, if the time ever came for you to implement a descriptive model, you would understand in the simplest way how it should work and the result you should yield.
The SQR3 method is a “reading comprehension method named for its five steps: survey, question, read, recite, and review”.
The purpose of the SQR3 method is to gather and remember as much information as possible from what you read. Reading is something you will likely be doing a lot when learning the more difficult concepts in data science. Therefore, you might as well ensure that you’re getting what you came for.
The five steps of the SQR3 method are to be completed in sequence:
- Survey: Gather information to prepare yourself for what you are about to read, including the title, introduction or summary, headings, tables or figures, chapter objectives, end-of-chapter questions, and any additional key features. This information will give you a foundation in what you are about to read and will give you an idea of what you should focus on, as well as what the goal of reading the item should be.
- Question: Develop questions using the key information gathered during your survey. This allows your mind to pick out key details to answer these questions as you go through your reading.
- Read: Answer the questions you have developed by reading. Now is also the time to see where your answers could have more information, or where you should develop additional questions.
- Recite: After you have completed reading a section or chapter, try to answer the questions you have written from memory. The key here is to not move on to reading the next section until you can recite answers to your questions easily from memory.
- Review: After completing the entire reading, go back over the questions you have written and see if you can still answer them from memory. If not, review your questions and answers to ensure that you solidify your understanding.
What this technique is good for:
Upgrading your reading comprehension to ensure that you understand and retain what you read.
How you can use it to study data science:
TowardsDataScience is one of the best online resources for data science articles and how-tos. However, can you with certainty say that you have retained everything that you have learned after reading the articles they publish? Furthermore, if an article is particularly complex and goes into the intricacies of mathematics or artificial intelligence, can you confidently explain what you just read in simple terms? The benefit of the SQR3 method is that it will help you improve your comprehension of each TowardsDataScience article you read. Not only that, but you will then be able to apply what you’ve learned to your own projects because you clearly understand what was presented. This technique is also beneficial when used to read and understand scientific journal articles by helping you focus on the important information and fill in your knowledge gaps.
The Leitner System was developed back in 1972 by German science journalist Sebastian Leitner.
The Leitner System is a study method that “uses flashcards, card boxes, and a spaced repetition scheduling system” that has been shown to improve learning and memorization. This method was one of the first study systems to use spaced repetition and has held on since with increasing popularity as many studies have proven its effectiveness. This method also involves active recall, and can therefore be considered a hybrid between both study methods (separately discussed above).
The system uses three flashcard boxes to facilitate spaced repetition:
- Box 1: contains new flashcards that are newly added and flashcards from box 2 that were answered incorrectly. Flashcards in box 1 are reviewed every day.
- Box 2: contains flashcards from box 1 that were answered correctly and flashcards from box 3 that were answered incorrectly. Flashcards in box 2 are reviewed every other day.
- Box 3: contains flashcards from box 2 that were answered correctly and flashcards from box 3 that were answered correctly (no changes). Flashcards in box 3 are reviewed once per week.
As you can see, the methodology is the same as spaced repetition that was described earlier and also uses the aspect of active recall to determine whether a concept needs to be revisited more frequently or less frequently.
While writing flashcards by hand can be time-consuming, digital flashcard tools such as Anki have modernized and sped up the process considerably.
What this technique is good for:
Regularly self-testing comprehension of topics.
How you can use it to study data science:
The Leitner System is best used when learning concepts in data science. For example, you could create a deck of flashcards that covers the different statistical analysis methods. This deck would include concepts from descriptive and analytical statistics with the name of the concept on one side and the description and an example of the concept on the other. By going through these cards using spaced repetition and active recall, you can rest assured that you will understand the concept and will be able to retain that knowledge going forward.
A mindmap is a visual representation of ideas and concepts as they relate to a single topic.
Mindmaps are great tools for connecting previous knowledge to new topics, as well as relating new information and concepts to a topic that you are currently learning. These meaningful visual representations can help you visually grasp the link between information. Additionally, creating mindmaps is a more engaging form of learning that helps you utilize information creatively and critically.
A study conducted in 2002 by researchers from Barts and the London School of Medicine and Dentistry found that mind mapping can increase retention by 10–15%.
To create a mindmap, write the name of a topic at the center of a page. Write keywords and concepts around the central topic to create branches of information. Then, draw additional branches to the keywords and concepts to provide more detail or to link additional supporting information.
What this technique is good for:
Relating concepts to one another.
How you can use it to study data science:
Mindmaps are excellent tools to use when relating different concepts in data science. Data science is complex, branching into programming, mathematics, data analysis, and data visualization. Mind maps can help you link the different areas together and understand the interplay between them. For example, you could discover how different data visualizations can be altered by the statistical methods used, or how machine learning and artificial intelligence can affect data analysis. Everything in data science is interjoined and mindmaps provide the ability to create graphical representations of the relationships found within.
Data science is a complex subject to grasp but it doesn’t have to be painstaking.
By developing intelligent study methods that help you work through material efficiently, it’s possible to learn the toughest concepts in data science. The trick is to use the right combination of methods that yield the most well-rounded understanding of a concept.
Key takeaways:
- The brain can only retain so much information, so it’s vital to revisit information regularly that you plan on needing in the future to ensure its retention.
- Active learning methods that force you to recall information and use information in new creative ways can help improve retention and conceptual understanding.
- Developing strong reading comprehension skills will ensure that you make the most of your time reading and that you will retain what you absorb.