Are “DIY projects” valuable or publishable?
You are asking whether you will be able to publish these data and use it as an edge. It all depends, as user2768's answer mentions, on the value that your work has. Specifically, you have to come up with a valid and important research question that you work will answer. For example,
the 5-year or even 10-year array of data for a number of sensors staying in your basement are useless
the online tool to monitor the dust quantity in the air in your neighbourhood is useful as a community project, is likely to be useful to your neighbours / town administration, and can be commercialized, but is unlikely to turn into a scientific article. I know of such a service in my city, and you can take some inspiration from it: Tion CityAir.
monitoring gas release and dust particle concentration next to a nearby industry plant or landfill is useful. It can become particularly useful if the administration of the plant / landfill is about to introduce some environmental safety measures and is seeking to quantify their efficiency. In this case, your research will be answering a well-defined question: "Are enter new safety measures here improving the air quality?". This is valuable and publishable. In this case, 5-year arrays of data are not required; most important is to monitor the air quality some time before the planned safety upgrade and some time after; I'd suggest at least three months before and after.
Bottom line: starting a 5-year data flow is easy. Coming up with a valid design of experiment is harder.
Specific points to account when planning an air quality monitoring experiment:
Are your gas sensors able to monitor the exact gases you are interested in? If you are targeting plant monitoring, ammonia and sulfur oxides come to mind, but there might be so much more.
What particular dust particles you are able to quantify? For the environmental safety applications, PM2.5 particles are often the most risk-inducing.
It is great that you have worked on such projects at such an early age and I wish you all the best for your physics degree.
You want to know whether there's value in your projects.
First, you need to consider what you want, perhaps in terms of commercialisation, impact, or ... That is, you need to establish what "value" means to you. Then you can establish (perhaps with our help) whether you should start "a 5 year study on Air quality & climate change using 10 arduino sensors (gas , dust , temp & humidity) and an arduino."
Secondly, you need to establish how you can achieve what you want. You've mentioned that you'll have breaks during your physics degree. (That detail has been edited out, see the archived version.) But, those breaks are short and you'll need longer to conduct your five-year study.
I think these points need further thought before you can decide whether results can be published and whether it is worth your time.
Can I publish such research anywhere?
I think it is better to consider whether you should publish, rather than whether you can publish. This has been dealt with elsewhere (it was written for mathematics, but it applies to physics too):
I would advise almost any junior high school student not to think about publishing their mathematical work. Note that I did not say to stop or slow down in the learning and doing of mathematics in any way. In fact, the point is that the publication process is something that is done by professionals largely for reasons of professional exigency and not because it is pleasant or educational in its own right. When I work with PhD students to try to get their first paper published, there comes a point where they realize that the amount of effort to do so (even after all the theorems are proven) is something like 2-10 times as much as they expected...Moreover, undergraduates who do summer math research are now being much more pressured to write up their results -- even when they are not really significant, and even when they were largely put up to the results by their faculty mentors -- and this is very worrisome.
The OP is significantly older, but I think the point holds.
The other answers for this question are good, but I'm going to follow a different path.
Besides the raw data you gather, whether relevant to others or not, you should document your process, including any hardware and software solutions you use or create.
If you get good data for yourself, others may want to gather that same type of data for themselves, after reading your blog (or whatever). You will gain experience in building, designing, and maintaining your equipment, which can be invaluable to others who may want to avoid all the trial and error you already have done.
You may be able to provide schematics/diagrams, parts lists, or even kits to these like minded people. Given enough time, effort, and interest, you may even find people who want to add to your project, growing it into a larger study.
With more people come more data points. What you are doing now might not be a research paper, but it could be if you have 1000 people/devices spread across a (large) section of geography. Given that they would be using the same designs (procedures, equipment, and software), it could be fairly easy to correlate the data into a something that actually means something to more than just your neighbors.
Even if that doesn't go anywhere, you might have a nice side business with selling the kits.