How I Overcame My Fear of Math in Order to Study Data Science
Fear is no excuse to give up on a dream, particularly if it’s all in your brain.
I feel uniquely suited to write this piece as someone who failed math in high school.
I didn’t know anything about data science until I went on Medium and came across the data science course Malaysia. It made me fascinated from the first piece I read.
Data science combines my passion for technology with business, engineering, and almost any other subject you can think of. As someone attempting to get into the environmental consulting field, I realised that data science may be the deciding factor.
I knew I had to at least give data science a go.
However, one criteria reduced my enthusiasm: a thorough grasp of mathematics and/or the capacity to teach high level calculus, linear algebra, and statistics to oneself.
I failed high school arithmetic, as I have said. I wasn’t getting off to a good start.
As I read papers outlining how much math you need to be a data scientist, I became more intimidated. Pre-calculus, calculus, multivariable calculus, trigonometry, linear algebra, differential equations, and statistics were all mentioned in those papers.
I’m not fundamentally awful at arithmetic, which is a strange reality about me. I’m simply not good at arithmetic that isn’t useful or applicable. I have no trouble dealing with math in statistics, physics, chemistry, or calculus. To tell you the truth, I’m really excellent at it. But I’m out of luck when it comes to more abstract applications like number theory and differential equations.
As a result, I was dreading a large amount of the mathematics I’d have to study in order to become this all-powerful data scientist.
The moment of discovery.
I thought the title was blasphemous when I first saw it. At the time, I assumed that anybody who stated that data science didn’t need significant math knowledge was merely searching for clickbait views.
How could one piece teach the polar opposite of all the well-received things I’d been reading? Was everything I’d been reading a bunch of nonsense?
As I read further into the essay, I saw that the author had drawn a distinction that I hadn’t seen in any other paper on the subject: there is a separation between theoretical and practical data science.
In a nutshell, the article illustrates how theoretical data science (which is often conducted by academics) differs from practical data science (which is usually practised by industry professionals). Academic data science is often significantly more mathematically intensive than commercial data science.
While the author acknowledges that the amount of math required varies based on your position as a data scientist (junior vs. senior), one of the most significant aspects is whether you will be working in academia or business. Of course, depending on the industry you choose to enter, you may need more math than in other fields. However, that level of arithmetic is unlikely to be as high as it would be if you worked in academia.
Further, fundamental data science skills like data manipulation, visualisation and exploratory data analysis do not need much math. Creating scatterplots or histograms does not need difficult mathematics, and may be done using high school arithmetic skills, according to the author.
When it comes to machine learning, one of the pillars of data science, the author debunks the myth that effective machine learning practitioners must have a strong grasp of mathematics. This is false, since the author discusses how numerous machine learning practitioners he knows do not have extensive math background yet work at prominent corporations like Apple and Bank of America. Those practitioners make up for their lack of theoretical understanding by their ability to use mathematical approaches.
There is also a distinction in the scenario in which you wish to practise machine learning, just as there is with other data science applications. The amount of math you need to know to develop effective machine learning models depends on whether you work in academia or industry.
The author contrasts reading an academic machine learning publication to looking at the results of a business machine learning model. It’s probable that the machine learning paper will involve difficult math, but the practical model will likely just need basic statistics, linear algebra, and calculus.
The author blasted open the doors for up-and-coming data science practitioners like myself who may otherwise have turned away due to a phobia of arithmetic by openly proclaiming that data science and machine learning approaches don’t need years of serious professional mathematical study.
So, what math abilities are required to begin studying and becoming a proficient data scientist? The author suggests focusing your studies on five easy topics of mathematics:
- Basic graphs and charts.
- Algebra is a fundamental skill.
- Statistics for beginners.
- Notation for basic math.
This set of mathematical abilities is more than enough for ambitious young data scientists to overcome any challenges. I’ve read many times how experienced data scientists value going back to basics and tackling issues the simplest way possible. Those straightforward solutions are most likely based on an Excel spreadsheet and a few basic algorithms. To put it another way, differential equations, complicated machine learning models, and artificial intelligence aren’t required for every data science challenge.
How is this affecting the way I study data science?
I wrote this post in December 2020 showcasing my new year’s resolve to study data science in 2021. In that post, I discussed the self-taught learning programme I had devised. Those publications provided laundry lists of mathematical abilities that anybody interested in becoming a data scientist should possess.
This article has made me rethink my degree and focus on more practical skills like descriptive and inferential statistics, basic algebra and data science principles. In a nutshell, the mathematics element of my education will resemble this:
- Mathematics with a limit
- Statistics: Descriptive and Inferential
- Linear regression, logistic regression, support vector machines, decision trees, are some of the algorithms and techniques use.
In addition, I’ll be concentrating my efforts on studying various data science and machine learning methods right away. However, I’ve learned from my prior math education that merely knowing how something works is adequate. This practical approach to arithmetic learning will cut through the clutter and simplify my grasp of the most crucial information. It will eventually be time to study why, but until then, I can get by with only knowing how to apply the method correctly.
Basically, I’m concentrating on the most crucial ideas. I’d want to concentrate on data cleansing and manipulation, data visualisation, and exploratory data analysis. I will be competitive in the industrial arena if I concentrate on the four basic data science abilities, not as a data scientist in title, but as someone who can use data science ideas to solve issues.
Finally, some thoughts.
As someone who has previously been put off by a lack of confidence in mathematics ability, I believe it is critical to examine any field from all perspectives.
In data science, everyone is so obsessed with becoming a “data scientist” that they learn difficult mathematics. Less math-phobic should master the practical aspects of their employment and apply them to occupations outside of the “data scientist” arena. People who learn new technologies often get promoted at work simply because that can benefit the company.
In order to become a data scientist, math phobe must undergo months of instruction in a subject they seldom comprehend. We can make an impact by grasping data science concepts, whether or not we can solve differential equations.
Explore more useful articles at articletab