Case Study
BeerBoard – Algorithmic Line Cleaning
Client
US-based Beer Management System
Technology
Have a similar project in mind?
Business Challenges
The flow meter attached to each beer tap is measuring the beer line cleaning liquid also as the beer poured, which is leading to inaccurate insights of beer poured and sold.
Our Solution
Helenzys proposed to build a reward and penalty based Machine Learning algorithm which will identify line cleaning liquid and automatically mark the same in the Database.
Salient Features
- Algorithm integration to control panel
- 85% prediction accuracy and can be further improved by training the algorithm
- Data collection
- data wrangling base version of Spike detection algorithm.
- Identify relevant features from the data set of Smart Bar and enhance the algorithm for higher accuracy
Product / Service Description
Initially, the line cleaning data was manually deleted by the L1 support team, which is an expensive and non-scalable solution. To overcome this, we have implemented automatic detection of Line Cleaning using the data available in Smart Bar with 85% prediction accuracy. As with any program implementing machine learning algorithm, the results will come with a confidence level. Build with Python for robust machine learning algorithm & deployed on AWS for scalability.
Result
Highly accurate beer monitoring insights
The client wants web application for bar admins to track the bar performance and work order details of their respective
The client synthesizes Poured and Sold data available in a bar and give insights for their customer through SmartBar app. Due
The client wants to completely redesign their mobile application with certain feature enhancements for iOS and Android
