I am posting today to share that the Python Developers Survey 2024 is now open! The PSF partners with JetBrains to run this survey to understand the current state of Python development and the community.
We encourage you to contribute to our community’s knowledge by sharing your experience and perspective. Your participation is valued! The survey should only take you about 10-15 minutes to complete.
We aim to reach even more of our community and ensure accurate global representation by highlighting our localization efforts:
The survey is translated into Spanish, Portuguese, Chinese, Korean, Japanese, German, French, and Russian. It has been translated in years past, as well, but we plan to be louder about available translations this year!
To assist individuals in promoting the survey and encouraging their local communities and professional networks we have created a Promotion Kit with images and social media posts translated into a variety of languages. We hope this promotion kit empowers folks to spread the invitation to respond to the survey within their local communities.
We’d love it if you’d share one or more of the posts below to your social media or any community accounts you manage, as well as share the information in discords, mailing lists, or chats you participate in.
If you would like to help out with translations you see are missing, please request edit access to the doc and share what language you will be translating to. Translation into languages the survey may not be translated to is also welcome!
If you have ideas about what else we can do to get the word out and encourage a diversity of responses, please comment on this thread.
We hope that highlighting these localization efforts boosts our efforts to capture a representation of how Python is used across the ENTIRE globe!
Regarding the translation errors: if there are just a couple of fixes, you can add them to this thread as a reply. If you have found many, I would suggest sending them in an email to psf@python.org.
Regarding the prizes: I understand. I’ve learned over the years that running giveaways is complicated. I’ve also learned that sending physical items (i.e. a t-shirt or mug) brings its own complexities. I am not running that aspect of the survey so I have little control over the prizes but I can pass along the feedback, and let’s see if it can be updated for next year.
One solution would be to switch from “X/Y/Z” gift card rewards to charity rewards, such as: if you answer this survey, 1 USD will be donated to a charity of your choice among X/Y/Z.
I translate survey kit to Persian(Farsi) but I found, it will not affective if all those questions being in English.
So maybe for next years we could find a platform that support i18n within survey and don’t messed with resalts.
Are these questions really representative of what the PSF want to know?
I see a couple of boilerplate experience and background questions then the rest is almost exclusively what IDE you use, why you use that IDE, how you use your IDE, what tools/frameworks/databases/AI assistants/… (most of which PyCharm happens to have corresponding paid features for) do you use, what IDE features do you rely on…
There’s not really anything about Python itself, not even a mere do you like using Python? There are plenty of debates that happen on here which could be better grounded if we had real data on questions like whether people think Python and it’s packaging ecosystem is changing for better or worse or just moving the goal posts around, if it makes (breaking) changes too fast or too slow, should it prioritise being faster/smaller/lighter/richer/more expressive/more batteries included. It would also be nice just to cast a wider net on questions like what you think Python’s (ecosytem’s) biggest pain points are (with a nice big checkbox next to xkcd 1987 for me to click! ).
I’m trying not to sound conspiratorial but this current survey feels like marketing data for JetBrains rather than something to benefit the Python community.
Some of the multiple choice questions seem a little misguided. By tossing GUI frameworks, ML libraries, and data validation libraries (and only 2 of a much richer ecosystem of a couple of each of these) under the same category, it’s only going to validate what the survey asker seems to be aware of (at least, taking into account the issues with handling “other” in surveys)
Thanks, all, for the feedback- keep it coming! Plus points for giving concrete suggestions for questions we can add
I am not sure how many changes we will be able to implement for this year’s survey now that it has been launched but we can request changes for next year’s survey!
At the risk of saying something obvious, some light feedback for next year’s would be:
engaging in this “user success working group” which sounds hella cool: User Success WG status
brainstorming questions relevant to the various topics here (e.g. a section on typing (although there was that typing survey), about async, about packaging, etc…)
looking at PSF mission/goals and mapping those to questions (I bet more questions around diversity and representation would be very useful)
I liked the shortness of it. I’m never quite patient with polls like this but I didn’t feel this one dragged on for too long. That’s a plus in my book.
I didn’t quite grasp where the poll was trying to take me. Or more accurately, I couldn’t quite figure what the PSF was trying to answer which made it a bit more complicated to engage.
This is indeed a recurring concern among the makers of the Python packaging that there is a lack of concrete data about what users of the Python packaging ecosystem truly want, need, wish, like, dislike, and so on [*]. So, indeed, maybe it would be a good idea to sync with them when preparing the next survey. Just a suggestion, I have no idea if it would be indeed a good match, but maybe worth an investigation. I am not the right person for this, but I can at least try to find the references if needed.
[*] Besides the usual “Python packaging sucks and XKCD1987” that bring nothing but lowering the morale of the makers of Python packaging ecosystem. No worries, I saw the smiley, but still…
I agree that many of the questions seem pretty scattered, with heterogeneous collections of available choices. In one case “Scikit-Learn” and “Sklearn” were listed as separate options although those are just two names for the same thing.
I also agree that brainstorming the survey creation here or in a similar forum would be useful. Can you share anything about the current survey design process?
The other, perhaps more controversial suggestion I would have is to have more questions that are about opinions rather than “facts” like “what IDE do you use”. There are quite a number of issues on which, shall we say, distinct viewpoints seem apparent in many discussions here. Packaging is a prominent one, but also (just off the top of my head) type hints, backwards compatibility concerns, the usefulness of a “batteries included” library in an era of widely available fast internet, etc. I think it would be valuable to leverage an annual survey to take the temperature of the userbase on such issues[1]. This could be done by having sequences of 2-3 related questions on each of a few such topics.
As an example, on this survey there were one or two questions about what version of Python is used and how/when it is upgraded. But it would be useful to have responses to questions like “How often do you refrain from upgrading Python due to concerns about backwards compatibility? (Multiple times a year/once a year/less than once a year)” or “How do you feel about the level of backwards compatibility maintained from one Python version to the next? (Python should move faster to remove cruft and support modern workflows, even if it means sacrificing some backwards compatibility/The level of backwards compatibility seems just right/Python should prioritize backwards compatibility more, even if it means delaying potential improvements)”.
The other suggestion I would have is that it would be nice if a form of the data as raw as possible (i.e., anonymized but not aggregated in any way) were made prominently available. This allows third parties to slice and dice the data in different ways than whatever may show up in official presentations of the results.
This survey leaves me with a terrible taste in the mouth every year (I have discussed this with a few people like @Mariatta who might be better to explain how I feel in words).
For a long time scientists have known that self-selection bias in surveys is a huge problem. For me, as a non-native English speaker, “developer” usually means “professional software developer”, I’m not a professional software developer, this survey is not for me. Maybe we need a well designed and not IDE seller provided “Python Users Survey”?
A small example:
For me, teaching Python for years, this question made no sense and didn’t have a “doesn’t apply” option. You see, “professional coding experience” sounds like something “professional developers” have. I’ve been told in English this isn’t what it is meant. But even so, how about people using Python for hobbies? Could it be phrased something like: “For how many years have you been using Python for work or other projects?”
Scaffidi et al. in 2012 showed an estimate that in the USA there were more than 4 times the number of end-user programmers than professional programmers (a.ka. developers). Is the PSF survey reaching a significant sample of us, end-user programmers?
Ah yes, that one got my eyebrows wiggling too. My own Python journey consists of 2 years of hobbying, then a few years in academia, then I wandered into a (relatively high profile) open source project for a couple of years before finally getting a professional job with the word Software in its title. I remember 3 years back, just before the professional job, doing this survey and acknowledging that, at a time that I was pushing releases onto PyPI with ~1M/month downloads, I was doing so with “zero years of professional experience”…
(I personally just pretend the word professional isn’t there and give the survey the number of years I’ve been using Python. It feels to silly to give the literal answer.)
Did feel too much like just collecting market research data for JetBrains by the end, feels a bit odd that this is also apparently the official PSF Python survey
Hi folks, thanks for all the feedback. I am seeing a lot of actionable changes/ideas, which is very appreciated.
In response to @BrenBarn’s question regarding the current survey design process: we have a chance to review the questions in advance (PSF Staff, PSF Board, & other community folks we invite) and we can suggest edits or additions. I’ll be frank here, we don’t have data experts paid to work on this survey or PSF Board members directly sponsoring the survey in a hands-on manner (though some of our staff and Board members might be great with data, that is not their $dayjob, and board members can choose where they focus their time). My role is focused on community/communications- so as a non-coder, I rely on Python experts around me to review and make pertinent suggestions. That’s why I’m especially pleased to see all the feedback on the thread this year!
This brings me to the topic of JetBrains running the survey- I’m hearing that folks don’t love the idea of providing free market research for JetBrains, which, makes sense to me. I am sure they are getting something out of it. On the other hand, if they didn’t run the survey for us, it probably wouldn’t be run. Multiple folks at JetBrains are doing this as part of their day job- and I think we can generally agree that we want corporations to contribute to Open Source projects with more than $$. As I mentioned, the PSF doesn’t have anyone dedicated to running this (we are only 12 people, doing a lot!).
I’ve seen a handful of dedicated blog posts on the most recent survey results(1, 2, 3, 4), I’m sure there’s more I haven’t seen, and I coordinated a live stream event with JetBrains to discuss the results (which has 6K+ views). Based on all that, I’d like to think the survey is useful beyond market data for JetBrains. We are thankful for their support & labor (this is the eighth year they’ve administered the survey!).
The questions I will put forth (and some folks are already responding to) is, what would make the survey useful for you? What do you want to know about the current state of Python? I plan to go through all the feedback and make actionable suggestions to the JetBrains survey folks to improve for next year. I can’t promise that everything will be implemented- but I will push for changes to make it as useful as possible for our community
Contribute to the Python Developers Survey 2024 
).
