7 Data Science Trends You Need to Know Going Into 2023 7 Data Science Trends You Need to Know Going Into 2023
2022 was a year of notable advancements in data science and artificial intelligence across the board. Hardware became more accessible, AI... 7 Data Science Trends You Need to Know Going Into 2023

2022 was a year of notable advancements in data science and artificial intelligence across the board. Hardware became more accessible, AI hit the mainstream, more healthcare developments found practical applications, and more. As we begin a new year, here are some of the top data science trends we saw dominate 2022 that we can expect to be prevalent in 2023.

Generative AI and AI Art

There was nothing on our radar more this year than generative AI and AI art. Social media users used apps like Lensa to make AI-generated content for their online profiles, entertainment companies found practical uses for deepfake videos and voices, and in general, AI has hit the mainstream. Even non-AI and data science professionals are talking about the benefits and controversy surrounding AI.

One prime technology to come out of this is Stable Diffusion, a text-to-image diffusion model that can generate photo-realistic images with any text input, similar to DALL-E Mini which was popular over the summer. The technology was a bit controversial at the start, though feedback online has led to them making some changes regarding the issues surrounding copying artists’ work.

Data-Centric AI

We’re done with using whatever data we can get and hoping for the best. This past summer, Andrew Ng, a world-recognized leader in AI, proposed “Data-Centric AI,” which he describes as “the discipline of systematically engineering the data needed to build a successful AI system.” In simple terms, data-centric AI is meant to improve the funnel of data, ensuring higher-quality data for better outputs, rather than using whatever data is easiest to gather.

The term may be fairly new, but the practices have been around a while – there’s just now a greater focus on taking data seriously. There are a few things you can do to implement data-centric AI, such as data augmentation, feature engineering, and better-employing domain knowledge experts to ensure the relevance and accuracy of data.


The Fall and Rise of Hiring

2022 was a precarious year for job security. Some of the biggest employers in the world, like Meta, Amazon, Twitter, and others, made headlines with mass layoffs across the board. For social media industries, decreased user growth and erratic behavior of decision-makers led to a lack of job security, whereas plummeting sales affected job growth in retail and e-commerce sites.

However, not all hope is lost. Startups and mid-sized companies are rapidly hiring data science professionals, while larger biotech, IT, and finance companies are still comfortable fields for data scientists.

The End of Data Scientists?

No, this isn’t as grim as it sounds, especially given the last point. By now you’ve surely heard the phrase “Data scientist is the sexiest job of the 21st century” countless times, but now it’s not so broad. Other job titles, like data engineer, machine learning engineer, and so on are on the rise. If you’re a career seeker, find a niche and explore other titles as opposed to just “data scientist” and you may find something not just better suited for you, but something with a better chance of landing.

Large Language Models

Similar to generative AI and AI art, large language models entered the mainstream a bit, though maybe not to the same extent. Tools like GPT-3 can converse like a human, as they’re trained on absurd amounts of parameters. In GPT-3’s case, 175 billion parameters and 570 gigabytes of text.

These tools are starting to find real-world applications, finally branching out of academia and research-only initiatives. Businesses are starting to use GPT-3 and other tools for things like chatbots, question-answering systems, and so on. There’s still a need for human operators, however, as some of these systems, like Meta’s BlenderBot 3, still have some kinks to work out.

As a late entry into trending AI news among these data science trends, OpenAI unveiled ChatGPT in November, and it’s certainly a topic of conversation. While it’s still new, it’s proving to be quite powerful, as it can answer questions and even write essays, causing headaches for those in education.

Data Governance and Responsible AI

As artificial intelligence and data science trends continue to permeate pop culture, social media, government, and business, and make news headlines across the globe, more attention is being paid to the ethical side of AI.  AI governance is a term we a heard a lot this year, especially as AI-based controversy seems to be a hot topic. Organizations – from government to startups and everything in between – seem to be more transparent about their use of AI, use better data, and focus more on the diversity of data. 

AI governance is sometimes synonymous with responsible AI, as both terms focus on the ethical use of AI in practice. Businesses are appointing individuals whose sole purpose is to ensure that their AI practices are ethical, and even the US government has unveiled an AI Bill of Rights.

In 2023, all researchers and practitioners should work on their own AI governance processes, such as transparency, a lack of bias, and so on. No organization should want to deal with the fallout from proving to be using unethical AI!

The Need for Real-Time Data

People, trends, and habits constantly change. A person’s shopping habits will change around certain holidays, so organizations shouldn’t be making decisions based on outdated summer data. Real-time data, aka streaming data, isn’t optional anymore, and any organization that wants to take itself seriously should implement a process to gather data in real time.

This isn’t just about optimizing prices as something that an e-commerce company would do. Real-time data can also help with risk protection to guard against threats as they happen, making the goal prevention instead of recovery from malicious attacks. Gathering data as it happens is a huge component of machine learning safety, and is considered the future of cybersecurity by some.

How to Stay on Top of These Trends

There’s a lot for an individual, a team, or an organization to consider among these 2023 data science trends. Staying on top of the news and going about your daily work won’t cut it anymore if you want to get ahead or even stay relevant. With ODSC, there are a number of ways that you can not only stay up to date with all of the 2023 data science trends mentioned above, but also be prepared for future data science trends as they appear.

  • ODSC East 2023: A 3-day data science conference in Boston, MA, this May 9th-11th, featuring talks, workshops, training sessions, and networking events – all designed to keep you current with all things AI.
  • Ai+ Training: An on-demand platform for data science training, featuring past ODSC sessions, on-demand talks, and live sessions each month. A yearly subscription gives you access to everything the platform offers and even exclusive ODSC event discounts.
  • Data Engineering Live Summit: Coming up this January 18th, this free virtual event is devoted to all things data engineering, with talks focused on topics like data pipelines, ETL, building applications, and more.
  • Ai+ Careers: Looking for a new job? Check out our Ai+ Careers platform where you can upload your resume and get matched with the jobs best suited for your skillset.

Alex Landa, ODSC

Alex drinks a lot of coffee and sometimes writes stuff. Connect with him on Twitter @alexandermlanda or LinkedIn: https://www.linkedin.com/in/alexander-landa-a0889857/