Another little side project of mine bites the dust, but not due to failure. I discontinued this one due to hitting the limits of my hosting resources and a few other reasons. Read of if you're curious.
What was hubreports?
If you have no idea what hubreports was, I'll try and provide a simple explanation. It was a side project of mine that accessed GitHub's API on a daily basis. Information was collected on each programming language (that GitHub recognised): Counts of repositories / users / code, top repositories and users, top new repositories and a few other bits and bobs.
From all the information gathered, several reports were generated as a website. From the website you could browse these statistics for github as a whole or a specific language (some features were restricted to the top 10 or so languages). Historical information was also retained, so you could see how a repositories number of stars changed over time or their chart position in relation to other projects.
Why create it?
Originally I wanted to keep track of new popular repositories on github for my own blog posts. But the more I dabbled with their API, the more I wanted to create something larger out of it. I also wanted an excuse to put Node.js and MongoDB to good use, provided a good experience to gain knowledge with them.
Bad design choices
I used hubreports as a learning exercise. It became too easy to pick various new trends and try them all out at the same time. Looking back, I should have stayed well away from AngularJS for this project. Without having the spare time to come up with a solution, it meant that the site wasn't going to be indexed by search engines correctly. With few ways to encourage visitors, it ended up getting pretty quite as no one knew it was there.
No budget ;)
It was a personal project, so with no relation to my day job there was never going to be any help forthcoming with any costs from there. Personally, I've got zero money spare for this sort of thing. On the odd occasion with my little pet projects, I'll splash out on a domain or some other cheap one off cost, but any regular cost would always be out of the question.
Thankfully there were resources available to help with this to a degree. I took advantage of OpenShift's free tier to host the project and it worked perfectly...
End of the line
Everything has a limit, especially stuff that's offered for free. Eventually hubreports had sucked up enough data to reach the memory limits on the OpenShift free tier. At that point the database server was no longer responding to any requested and I struggled to even get my data out of it manually. That meant that I could either try and slim down the database or call it a day.
I still consider the project a success. Sure, it didn't make me millions, but it did help my wrap my head around Node.js and move into more interesting areas with it. From the knowledge I gained, I felt confident enough to speak on Node.js and Express at a conference for the first time. Then off the back of that I worked with Matt Gifford on three Node.js related articles for Web Designer Magazine.
It's a success because all I want out of my side projects is a learning experience. Sure if one of them ever makes me some money or secures me an amazing job somewhere, I'll be really happy. But that's never the reason I start any of them ;)
There's something I'm playing around with. It'll be GitHub related again, simply because I love fooling around with their data and I'll also make use of OpenShift again due to my limited resources. I don't want to give much else away but it'll be a bit of useless fun and not very productive at all. My challenge with it will be to try and make use of Web Sockets and possibly a client side framework like React or something similar. Just don't hold your breath that it'll be out any time soon ;)