Hey folks! Let me share this here. I think it is not yet.
Julia is going upwards in the Redmonk rank, which means that we are approaching the diagonal while we have more topics in StackOverflow. It’s clear that, because of the open nature of the language, we have had a good basis of projects on GitHub. The new trend shows that we have now more users, and that’s encouraging!
I’m personally excited about the release of 1.0, which would attract even more users on the following course. Congratulations to all of you, and specially to the core developers of this language and package developers!
Nice to hear a good news!
I’ve added Redmonk ranking of Julia on Q1 2017 at Q1 2018 graph
That’s the way of trusting news for a smart person
Don’t get too excited. If you look at the results from the latest kdnuggets poll you’ll see that Julia usage did not grow over the last year New Leader, Trends, and Surprises in Analytics, Data Science, Machine Learning Software Poll - KDnuggets
Hey, this message maybe is OK for you, but I get as excited as the situation excites me. What you send indeed is more related with machine learning, I guess. And that of course requieres more production ready tools as you may know. Julia is evolving yet, but is close to freeze its syntax so let then the packages mature.
Are you really following the development of Julia and are you using it?
I think we are doing ok, but I’m not going to get either excited or worried about these popularity contests for quite a while.
I agree with your skepticism, but I do like this analysis better than some others. This one actually presents its metrics in a way you can compare. It’s very easy to look across the axes of the graph and see that we’re approaching Mathematica in our number of SO tags, but we’ve well surpassed it in terms of GitHub repositories. On that front we’re in the same neighborhood as Clojure and Haskell.
And the metrics aren’t themselves meaningless — they do represent (in a limited fashion) how big the corpus of open source code is and how likely you might find a question about the language to have already been answered.
I also think there is something to be said for the quality of code available for Julia, aside from just the quantity. DifferentialEquations, JuMP, Flux all seem to go beyond what is available or even possible in other languages in a significant way. There are also some cool tools available such as StaticArrays, RigidBodyDynamics, Query which are certainly possible in other languages (and indeed have many other implementations) but have impressively straight-forward and succinct code in Julia. Lastly, a lot of the “core ecosystem” stuff like DataFrames, JuliaDB (which I expect to take on a larger role after 0.7) are starting to look pretty slick and offer some flexibility that is hard to come by in other languages.
10 years ago if you had shown me this stuff, I would have been blown away. Suffice it to say that’s very different from the puzzled reaction I had when I first saw Python.
I follow it from a distance and I did implement a recommendation system several years ago in Julia 0.3. In the mean time I backed off from further use and went back to R for data analysis.
It would be interesting to see the 2017 and 2018 graphs superimposed, 2018 in black, and 2017 in red, so that the movements of all languages could be seen easily.
It’s a bit hard to see. here is superimposed graph:)
If the scale is off in this way then all languages will tend to move upwards and to the right from 2017 to 2018, so Julia’s growth in the rankings is somewhat less than it seems here.
Yes, I also noticed that quite a few languages that I don’t think would have moved at all, made similar moves, like
Yeah, It looks weird Q1 2017 graph didn’t have
100 tick on their axis. something must be wrong.
Q1 2016 graph seems fine. So here is superimposed graph with Q1 2016 ranking
For me one of the more promising signs of growth was Julia’s presence in the “most loved languages” part of the latest Stack Overflow Developer Survey. Not that this means a whole lot, but it marginally beat R, which I found pleasantly surprising.
Considering that many more people know about R than they do about Julia, this is a good sign at least.
I am not sure this translates into “growth” by any meaningful definition of the term. AFAICT the “love” index is simply the fraction of current developers who want to continue developing in a language. Growth would be in the denominator, but the statistic is about the ratio. As an extreme example, if only one person was using a language in their sample, but they wanted to keep using it, that language would be “loved” 100%.
Also, aggressively promoting a language would probably decrease this ratio, since it would bring in developers from the margin, a lower fraction of whom would have an interest in investing in the language compared to the early adopters/core.
The Most Popular Languages by Occupation helps put things in perspective. Julia simply does not show up, not even for “Data Scientist/Engineer”. This is probably because it is a niche language, aimed scientists who program, who are a tiny, tiny subgroup of all programmers.
I tried to import an Excel file in julia and it didn’t worked so I stopped learning it for sometime. I would be more than happy once the official 1.0 breaks in. Would be the first one to learn it.
I had some trouble with reading Excel files lately too. I wanted something with minimal dependencies (Taro requires JVM) and the FileIO implementation hasn’t been updated to DataFrames 0.11 yet (I know it’s been reported and will eventually change). In the mean time I had a surprisingly easy time with XSLXReader
I noticed that Julia was on both “most loved” and “most dreaded” languages lists on the 2018 stackoverflow survey. I haven’t seen the survey that these results are based on, but it almost seemed as if one of the questions was “rank all languages”, and most people, not knowing what Julia was, had it pretty far down the list. Like I’ve said before, these things aren’t very meaningful, especially for something as relatively obscure as Julia still is.
There seems to be an increasing amount of obsessing over this stuff here on discourse, which seems really unhealthy to me. It seems likely that we won’t really have a good idea of how well the language is doing until a year after 1.0. In the meantime lets make the most of it and enjoy that it exists.