By digging deeper into Stack Overflow page view data, one can uncover a unique advantage Google has in data science tools.
There are many different ways to assess the popularity of different technologies. Redmonk looks at GitHub and Stack Overflow data to measure programming language popularity. Ryan Williams pulls HackerNews mentions. DB-Engines scours a variety of sources, including Stack Overflow, to rank diverse databases. Among these different measures, Stack Overflow data is often a critical part, but perhaps not the Stack Overflow data that most pull.
Indeed, as Google's Felipe Hoffa shows, a better metric for assessing popularity of different technologies might be Stack Overflow page views, instead of new questions asked. Digging into the volume of page views yields very different results on what developers actually care about.
We visit what we love
It's possible, of course, that page views doesn't measure the popularity of a technology, but instead could reveal its deficiencies. As Scott Hirleman pointed out, "[I]t could be 'what's truly broken' if you were to compare usage to page views. If people have to search all the time, then the docs could suck."
Hoffa, however, had two good responses to this critique. The first is that whether a developer loves or hates technology, the fact that she's visiting Stack Overflow suggests that she's using it: "When users love something, they use it more. The more they use it, the more unexpected ways they try to combine it and extended it," Hoffa wrote in a tweet.
The second is related, and suggests that even eventual hatred of technology likely stems from an initial love for the tech: "Love brings usage, usage brings frustration, frustration brings hate. There's no hate if users don't love/use the product first. (At least when there's a choice)," Hoffa write in a tweet.
SEE: IT leader's guide to the future of artificial intelligence (Tech Pro Research)
Again, whether someone chose to use a technology, and whether she loves or hates using it, is somewhat immaterial to Hoffa's analysis—if they're on Stack Overflow trying to dig up information, they're using it. But what are they using?
Questions and views
Stack Overflow provides access to its data trends tool, but it doesn't measure page views. Rather, it measures new questions asked for a given tag, like TensorFlow. That's nice, but it leads to erroneous conclusions about what developers care most about, as Hoffa's example of Git versus TensorFlow indicates:
Git is obviously more popular. TensorFlow is simply sparking a lot of new interest.
Amazon Web Services (AWS) pays Google back in serverless, with AWS Lambda bullying Google Cloud Functions and Microsoft Azure Functions. But in the critical battleground of data science, using TensorFlow or Google BigQuery (versus Amazon Redshift), or other related areas, Google's technologies are really popular compared to Amazon's. That may not be playing out in their respective cloud revenue income statements (yet), but it's worth watching.
- Special report: How to implement AI and machine learning (free PDF) (TechRepublic)
- Google makes it easier to incorporate machine learning into mobile apps (ZDNet)
- Machine learning: A cheat sheet (TechRepublic)
- Google preps TPU 3.0 for AI, machine learning, model training (ZDNet)
- Cloud AutoML highlights Google's ability to build open ecosystems and monetize them(TechRepublic)