It's hard to explain the functionality of the app in words so here's a user experience flowchart of'Find My Train':
Hopefully, that gives you the gist of what we're aiming to do. This app was conceptualised from the following key problem statement:
Currently there is no app to look at, it's all one enormous thought baby but we're currently working towards our MVP (minimum viable product).
We're a small team of 4 aiming to build this product in a week. Fingers crossed!
To start thing off we outlined our key problem statement and brainstormed some features we would want to include in the app. The useful output from that session is shown below
"I'd say our research was thorough, in-depth and useful" (Absolutely No One On Our Team - 2017 )
Key focuses on:
- Which platform to deploy on?
- Which APIs for messaging, trains and geolocation already exist?
As a commuter
I want the app to find out the nearest station
so that I don't have to input it myself
As a commuter
I want the app to tell me what train I'm currently on
so that I can inform my friends
As a commuter
I want to tell the app where I'm going
so that it can tell me the estimated arrival time
As a commuter
I want the ability to send my travel details to my friend
so that they know when I'd due to arrive
As a commuter
I want the app to inform my friends of any delays to my journey
so that they know I'm delayed en route
As a commuter
I want...
- Node.js: Handy for single page web app deployment to Heroku
- Javascript: Goes without saying
- Jasmine: Unit testing
- Travis: Nah this one's a lie. Maybe one day the green build stamp will be there
- Geolocator: Inbuilt into most browsers, pretty handy really
- Transportapi.com: Pretty much holds up our entire app
- Whatsapp API: Spamming all your contacts with train info can be oddly rewarding.
It could be live on Heroku here but, to be honest, probably isn't. If you want to check this out locally follow these steps!
We're assuming you are on a Mac set up for development, are more competent than us and that you have Homebrew installed. If not the following links might help: Install Homebrew here. Set up for development here
This cracking installation needs Node. In your terminal run:
$ brew install node
$ cd file/location/of/train-spotter
$ npm install
$ node index.js
Check out the following into your URL bar in your browser:
http://localhost:5000/
I like trains. It was brilliant to work solidly on a week on a project of our choosing.
We tested the app on an actual commute! It worked! Getting the whatsapp integration and sending out actual messages was brilliant too.
It doesn't look half bad, although leaning on the design crutch that is bootstrap did help with that enormously.
All of us now feel comfortable with (and possibly even excited about) tackling API calls and writing asychronous Javascript.
Standalone API calls are fine but when you're chaining 4+ it starts to get a bit challenging, set back the project timeline by about a day. Thorough doc review and youtube research set us on the straight and narrow though!
As can be expected from any one week project we were too overambitious on the minimum viable product. The features we wanted to implement were slightly beyond our reach in the timeframe of 4 days. After day 2 we realised we needed to largely scale back the app in order to put a nice looking product!
'What's a feature test?' (Rupp, N 2017)
'Spike for life' (Odendaal, R 2017)
The first two days were spent wrestling with Asynchronous Javascript. This meant numerous overhauls of the codebase which left the testing somewhat to be desired. We didn't manage to get Travis working but to be honest, when you're working to a deadline it's hard not to be impatient with setting up that friendly code miner.
Our programme almost exclusively relies on the perfect user. The interface doesn't give feedback to the user to let it know what it's doing and won't let them know whether it's a slow API call or one that's just flat out broken.