The web is underpinned by JavaScript, which was initially created in the 1990s to enable webpages to be edited. Its popularity and range have since grown exponentially, making it a serious contender amongst the most widely used programming languages.
As its usage increases, many are questioning whether JavaScript is suitable for data analysis and data scientists. It is not a straightforward answer. Therefore, let us assess the advantages and disadvantages of JavaScript as a language for data science work.
An Explanation of Why JavaScript Is Bad
It is generally accepted that Python, R, Scala, and Julia are the most suitable programming languages for Data Science. However, it is not advisable for beginners to start with JavaScript as their language of choice.
In acknowledgement of JavaScript’s widespread popularity, the author of the book, David Beazley, states in the preface that he initially planned to give it the title of ‘JavaScript Vs. Data Science’.
JavaScript is well-known for exhibiting a range of quirks, particularly when dealing with numerical values. NaN (Not-A-Number) is a commonly used representation of values that cannot be represented using the numeric data type and was first introduced as part of the IEEE 754 floating-point standard.
When attempting to divide by zero using JavaScript, one will be presented with the result of ‘NaN’. This is not particularly informative, and what is more concerning is that ‘NaN’ is classified as a number within the language, making it difficult to identify the source of the error.
It is worth noting that, while this is a minor inconvenience, it does point to a broader issue; as JavaScript is dynamically typed, it has an imprecise way of determining whether a variable holds a numerical or textual value. Although this is not a major problem, it does necessitate the implementation of protective programming.
When dealing with large quantities, it can be difficult to manage. JavaScript is not suitable for working with high volumes of data, as it is not precise with larger numbers and does not support multithreading or parallel processing. Additionally, both JavaScript and Node.js can struggle to cope with computationally intensive, CPU-bound operations.
Despite the potential to mitigate certain issues, the opportunity cost involved may render progress unfeasible. A data scientist may have to consider whether the time invested into learning JavaScript is worth it, when there are other languages which can potentially provide the same results with greater efficiency and less difficulty.
Time spent learning JavaScript is time not spent learning other languages… It, though, may not be such a terrible thing after all.
JavaScript: The Reasons Why
Beazley’s primary assertion in his publication is that the issues pertinent to JavaScript have been thoroughly addressed, and that the interest of the JavaScript community in data science has grown considerably in recent times. Consequently, there are now resources available to make it a viable choice.
JavaScript’s ease of access and legibility make it an ideal programming language. As an example, if you are viewing this on a personal computer, simply pressing F12 will bring up the JavaScript console.
Gaining knowledge of JavaScript is simple, and its widespread popularity ensures that a wealth of resources are available to assist with the learning process. A quick glance at the figures available on StackOverflow demonstrates the extent of information that can be found regarding JavaScript.
It is a benefit that an increasing number of companies are producing their products using web technologies based on a Node-based stack. A data scientist’s capacity to communicate effectively with the product development team is improved by their shared language.
The widespread use of the same technology makes it easier and quicker to incorporate new products and services. It is analogous to having a conversation with someone who speaks your language.
Microsoft’s TypeScript is a superset of JavaScript which addresses one of the main criticisms of the latter language: its weak typing. TypeScript’s stricter typing system is even more stringent than that of Python, another popular language among data scientists. Static typing of languages leads to improved programming practices and fewer bugs in code.
Microsoft’s Napa.js is an attractive option for those who require support for multiple threads, yet it is still in its early stages and may not be the optimal choice. It is a testament to the appreciation and promotion of JavaScript as a versatile language, capable of being used for both data analysis and other purposes. However, this is not the only advantage.
Upgraded Data Science Resources
It is widely acknowledged that JavaScript does not offer the same range of data science libraries as more powerful languages such as R and Python. We fully agree with this view. It is clear that any aspiring data scientist must possess a range of additional skills, even if they are a strong advocate of JavaScript.
The use of JavaScript in data science has grown exponentially in the past five years, with the introduction of the TensorFlow JavaScript library being a particularly noteworthy development.
It is important to note that there is an emerging JavaScript data science landscape. D3.js is a renowned data visualization library that provides a comprehensive selection of browser-based tools for developing dashboards, reports and data storytelling.
TensorFlowJS is a powerful example of the potential of Machine Learning. TensorFlow, a popular Machine Learning library, has been implemented in JavaScript, allowing ML algorithms to be executed within the browser and/or Node.js server.
In response, I would query why one would wish to take this course of action. It is true that working on a web browser may not be ideal, however, it is a practical solution for prototypes, small projects and programmes that are not memory intensive. The time and effort that is associated with setting up a virtual world may not be justified when a web browser would be sufficient.
We now have the opportunity to utilize various tools to interact with the language which powers the internet and online applications. Browser-based data science offers us the opportunity to explore new ways of processing and presenting data in a user-friendly manner.
It is possible to access the results of a data narrative presented through a web application constructed with JavaScript, HTML and CSS, within a few seconds, providing the user has a smart device and internet access.
This rise in the use of JavaScript in Data Science further demonstrates the evolution of the industry. A data scientist is no longer just an individual who works in a solitary environment analyzing data; they are storytellers who need to devise innovative ways to communicate their discoveries and encourage data-driven organizations.
Simply Another Method to Use
It is clear that understanding JavaScript is not yet a necessity for data scientists, however, it has the potential to be a beneficial addition to their skillset. This knowledge may not revolutionize the field, but it has the potential to open up a variety of new opportunities. Therefore, it should be considered a positive development.
