An iOS and Android Application that recognises whisky labels
The core purpose of the App is to allow the user to scan a whisky bottle label. The App recognises this label and displays the relevant details about the whisky stored in a back end database. Along with showing details of the whisky such as region, tasting notes and description, the app shows the average rating of the whisky based on the community using Whizzky… this is very useful when you are at a bottle store you can scan multiple whisky’s and find the best one!
Social Media Integration
The user can login via Facebook giving them the ability to share their whisky collections with their friends. This also allows users to invite friends to use the App helping with organic growth of the App.
The App has a front end search function that searches the whisky database the search allows filtering by country, category and via keywords.
rating and saving
Users can rate and save their favourite whiskies for future reference. Ratings are visible to all users of the App which helps when choosing a whisky to buy.
As with any development project, the evolveIT team adopted an Agile methodology right from the beginning. The major corner-stone of this app is a well-functioning image recognition system, with this in mind we played around with a few technologies before settling on the final API we would use.
Experimentation is key when venturing into a new technology, and while developing the core functionality it was not uncommon to find numerous bottles of whisky scattered around our office with various members of the team scanning the labels with their phones - certainly a talking point with visitors to the office!
The design brief was to keep smooth lines, a simple user interface and a "classy" look and feel. Under the hood we needed to keep the code minimal and efficient, while catering for the advanced functionalities making use of the device's camera for the scanning of labels.
By keeping the code clean and efficient we ensure that the size of the application is not a hindrance for those wishing to download and install it.
speed and accuracy
The speed with which a user can scan a bottle label to the point where the system picks up the correct result is paramount. Keeping to the mantra of "speed and accuracy" we needed to fine tune the balance between server-side and on-device storage. Having the image recognition take place on the device itself paid off considerably - with 80% of labels being recognised almost instantly.
roll out and release
Being a feature App of a major whisky festival coming up, we had a tight deadline. As the iOS app store review process can take its time we decided to focus on the iOS version of the app first and then "inherit" as much as possible when developing for Android.
The tight deadline ensured a tight circle of develop-test-develop between ourselves and our client. The way it should be in any project of this nature.
Towards the end of the project we developed the Android and iOS versions of the app in parallel, which ensured an on-time release on the respective App Stores, in time for the Whisky festival.
As with most Mobile Applications, a major part of the system is a reliable back-end admin system. Whizzky is no exception. A complex back-end system allows the client to administer all whiskies present on the platform, as well as a reporting module. Added complexities included allowing the user to submit whisky bottles which could not be located on the Whizzky platform, giving the administrator the ability to add them to the system. Giving the user the ability to add whisky’s has the added advantage of getting the users to do the work of growing the database organically.
Together with the core development of the application we also created a system which would allow promoters at the Whisky event to encourage visitors to download and install the app. This system tracked the promoters and allowed them to see real-time statistics on their performance via tablet devices. The promoters were rewarded through various prizes based on key metrics. This method and system proved itself to be very effective and will be used at future events to drive downloads.
This is an ongoing project with a long-list of new features in the pipeline. As such we have ensured that the architecture is scalable and allows us to refine the code-base and feature-sets on an ongoing basis. We will endeavour to update this case-study over time as features are added.