Ah, another "State of" survey. Original, right? Cue the collective eye-roll.
But let's face it, 'data' doesn't exactly scream 'blockbuster thriller'.
Funny thing, though – it's this very unglamorous data that quietly runs the show behind every app, every click, and every digital interaction we make.
In our little tech bubble, it's all about the shiny new toys – the latest frameworks, the AI breakthroughs, the latest drama…Everyone buzzes about React, Vue, Next, Nuxt...But when did you last hear a fiery debate over SQL vs NoSQL? Exactly.
Data discussions aren't usually the stuff of legend. Or so we thought.
Truth be told, we half-expected our survey to be a swing and a miss. But what do you know? Responses started flooding in.
One month and 782 responses later, here we are.
Data's not dull; it's just been waiting for its moment in the spotlight. We're on a mission, not just to talk about data, but to turn the conversation on its head.
Buckle up – we're making data exciting again.
#1 - PostgreSQL leads the pack
It's the choice of over 63% of JavaScript developers, and for good reasons. Its appeal lies in its powerful ecosystem, exceptional JSON support, and a thriving open-source community.
This isn't just a preference; it's a significant trend.
It emerged as the number one database in the Stack Overflow’s 2023 Developer Survey, outranking the traditional stalwarts like MySQL, SQL Server, and MongoDB.
#2 - Relational databases are most preferred
More than 57% of respondents said that they relational databases over all other types.
Despite the buzz around NoSQL databases like MongoDB and Redis, the simplicity and structured approach of relational databases continue to be the mainstay. This preference highlights a trend towards reliability and familiarity in database management.
I even asked our CTO Rijk. Here’s what he postulated:
The most likely reason is that the vast majority of data applications have some sort of schema consistency need for type safety and expectability. In NoSQL, you can quickly end up effectively reimplementing a lot of the things that a SQL database already offers out of the box, at which point it makes more sense to just use a SQL database for what it's good at.
That being said, the true flexibility comes when you merge the best of both worlds, by either using a NoSQL on the side for the non-known schema data, or by relying on JSON fields which are becoming more readily available and powerful across SQL database vendors.
#3 - Developers are leaning towards hosted database solutions
Cloud is leading the way by far. 68% of respondents prefer hosting with AWS, GCP, Supabase, and PlanetScale vs. alternatives like:
- Managed DB services (RDS, Digital Ocean) (33%)
- On-premises servers (25%)
- DBaaS (15%)
This preference aligns with the broader shift in the tech industry towards cloud-based solutions and managed services. Developers want to focus on building and deploying code without the hassle of managing the underlying infrastructure.
#4 - REST is by far the most dominant API type
This one was probably the second-most shocking to us just because of how lopsided it is.
94% of respondents prefer to work with REST, as opposed to 43% preferring to work with GraphQL.
In a tech landscape that constantly chases the next big thing, REST's durability is a testament to its fundamental strengths in API design. We’re sensing a pattern here – tried and true solutions have staying power.
#5 - Cost is not the most important factor in hosting
Instead, developers prioritize performance, reliability, and ease of use. This finding reveals a significant trend in decision-making criteria, where function and user experience outweigh cost considerations.
Still, cost was indicated as A factor by 46% of survey takers, but only 24% noted it was the most important.
Taking this a step further, we also asked them what the primary factors of hosting should be. The top 3, in order, were
- Machine resource utliziation (RAM/CPU).
- Storage space.
- Number of read/write options.
The least popular way to price hosting? By charging based on the number of database records.
#6 - 64% of developers use AI for code generation
Remember when we said that other stat was the second-most shocking? Well, this one really threw us for a loop.
The biggest use case for AI in this survey was code generation, 20 points higher than the next answer - AI for recommendations.
Going a little deeper, we asked two questions
- If AI was too risky for data analysis to influence decision-making. On average, sentiment was generally neutral.
- When it comes to AI actually interacting with a database with CRUD permissions, overwhelming sentiment was agreeing that it was too risky.
What’s Next?
Think of this as the opening chapter in a much larger story.
In our upcoming series, we'll dissect every nook and cranny of our survey. We're talking a deep dive into each response, breaking it down across different organization sizes, regions, and levels of experience.
It's one thing to know what the trends are; it's another to understand why they vary from a startup in Silicon Valley to a seasoned developer in Berlin.
Stay tuned, and subscribe to our newsletter for the full series we’re calling “Data Drops”.