A Simple Method to Detect In-Demand Tech Skills

Ruan Chaves Rodrigues
4 min readMay 31, 2023

It’s no secret that the tech landscape is dynamic and ever-evolving. New technologies are born, they mature, and then, often, they are supplanted by something even newer and shinier.

As job seekers in the tech industry, it can be tough to keep up with the hot skills du jour. While there’s no shortage of articles outlining the most in-demand skills, few provide insight into how these lists are compiled or how you can track these trends for yourself.

Today, I want to share with you a straightforward method that, although not without its limitations, provides a surprisingly effective snapshot of the current state of demand for different technologies and skills in the tech industry.

The method can be boiled down to this simple equation:

Job postings (which represent demand) divided by Stack Overflow questions (which represent the supply of developers).

This ratio can serve as a fairly reliable indicator of how much demand there is for a specific technology or skill set in the tech market relative to the number of developers who are actively engaged with it.

Let’s walk through an example using my personal area of interest: Natural Language Processing (NLP). To get started, I searched for “Natural Language Processing” on Indeed, focusing on remote job postings. While scrolling through the results, I noted the keywords that frequently popped up.

The next step involved searching for these specific keywords on Indeed to see how many job postings featured each term. The objective here was not to delve into each listing, but simply to count the number of postings. Following this, I divided each keyword’s job posting count by the number of related questions tagged on Stack Overflow.

For instance, if you search for “Natural Language Processing”, you’ll likely see “AWS” mentioned in several postings. A recent search revealed 15,882 remote job postings on Indeed that mention “AWS”.

Meanwhile, there are 152,405 questions under the “amazon-web-services” tag on Stack Overflow.

Dividing the job posting count by the question count gives us a score of approximately 0.104 for “AWS”. This score serves as a proxy for how in-demand “AWS” is in the job market.

Applying this method to a range of keywords related to Natural Language Processing allows us to construct a list ranked according to these scores. This provides an at-a-glance view of which skills and technologies are most sought-after in the current job market.

The full list is featured here. Here is a summary of the conclusions:

  • Strong Communities: Some technologies may feature high on the list simply because they have strong Q&A communities that are often used in place of Stack Overflow. Two examples are Splunk and Tableau. Splunk has 130K questions on its community, while it has only 2K questions on Stack Overflow.
  • Top Technologies: Technologies related to MLOps , Business Intelligence and Data Analytics are ranked quite high. If we consider only technologies where Stack Overflow receives about the same amount of questions as the official forums, the top scores belong to Looker , Kubeflow, the ELK stack, Azure Synapse and Databricks.
  • Emerging Technologies: Despite being quite new, LangChain appears with a higher score than its competitor Hugging Face and established technologies such as Docker, Spark and Apache Airflow.
  • Least in-demand Technologies: General-purpose technologies related to web development ( Django, Flask, FastAPI, Spring Boot ) appear low on the list.

While this method is not without its flaws, it does offer a simple, DIY approach to gauging which skills and technologies are in high demand. This can be invaluable for tech professionals seeking to stay competitive in a rapidly evolving industry. As you plan your career development and decide on which new skills to learn, keep this method in mind to stay ahead of the curve.

--

--

Ruan Chaves Rodrigues

Machine Learning Engineer. MSc student at the EMLCT programme. Personal website: https://ruanchaves.github.io/