Like Jambo, I highly recommend you go for it. It is a great learning experience that you'll come to draw upon again and again. No matter what tech-stack you decide to learn first.
Learning the basics of building a RESTful HTTP service is the cornerstone of just about any data-driven app. Mind you, it is a big field with quite a few gotcha's for new players - especially when you venture into the area of security and account control - not to mention building for scale and flexibility. But that is for later.
Right now, I am assuming neither is a problem. And unlike Jambo and Batvink, I'd not recommend PHP, as it isn't an environment in and by itself but is reliant on a host of dependencies to get anything useful done. Also, getting to grips with an SQL database may be overkill for such a simple project and limited traffic. Plus, it is yet another dependency to bring on. Which can be translated to having the need to dockerize the project to ensure that the dev-environment is cloned exactly to the server environment to avoid any nasty surprises as updates and patches comes rolling in. Then we're suddenly deep into linux devops and deployment land, which can not be said to be beginner-friendly.
I am a big believer in keeping it simple. The less moving parts, the better. The less dependencies, the better. The less different environments to jump in and out of, the better.
Which is why I'd highly recommend using Go. Google "golang restful api" and you'll find plenty of examples on how to build a simple server. With no external dependencies. Just one binary that you can upload. Go is easy to learn, and there are plenty of tutorials, videos and free e-books available to get you started. And once you're hooked, you'll miss all the features that make life so easy in that language. Not least of which is the tooling. After installing Go, get VS Code and the Go extension for it, and you're up and running in no time. If serious with programming (in particular programming server-apps), I'd recommend also changing over to Ubuntu 16.04 or at least if running Win10 and not wanting to experiment too much with your current set-up, get the Windows Subsystem for Linux Ubuntu 16.04 BASH installed.
As for keeping track of clients, you can either get the device uuid, hash it using SHA1 and send to the server (never send this unhashed, as Google and Apple will get very cross and not allow your app onto their stores if/when discovered) and use this as the ID, or have the user create an account where the server assigns an ID.
Then on client first-connect, search existing IDs, if not found, create a key-value pair where the ID serve as key, and the current date is the value. This get appended to a flat text file (preferably JSON encoded, but most any encoding scheme will do). The entire list you keep in memory for speed and efficiency. If we're generous and say each entry is 100 bytes, then in a megabyte of RAM, you can store over 10.000 client ID/first connect date pairs. Even a Raspberry Pi (and cheap cloud VMs) come with 1Gb of RAM, where at least 800mb is available after the OS have gobbled up its' share. Anyhow, the only time you should read from the list-file, is during start-up of the server.
I would not bother with any database stuff until either you got several thousand users, or you expand upon the API to also keep and handle more types of data that would benefit from the relational goodieness of SQL queries. And even then, I'd only use the database as storage, and cache everything (well, what is most used anyway) in an in-memory key/value management system like Redis. When scaling, anything that hits the disk - or worse, an external resource like a common DB serving several API endpoints in a distributed system, each SQL query comes with a significant cost in response-time.
Thing is, as long as the external API remains the same, taking the same requests in and providing the same data out, how you go about handling it behind the scene so to speak, doesn't matter. You scale and optimize that depending on demand. You don't fret it doing complex stuff to handle huge traffic before you need to.
Hmmm, this suddenly became quite a long post. Writing server apps is what I enjoy the most, so I'm easy to get into a rant on the subject. It's nice and pure, with none of that silly UI/UX headache or compatibility issues to worry about. Just nice well behaved text under one environment that isn't liable to change under my feet