Executive Summary
Elliptica wanted to develop a modern forecasting solution to improve its accuracy when predicting time-series data. A solution was necessary, as Elliptica had many client opportunities involving time series forecasting. Elliptica would need a data science team to create algorithms within the customer’s AWS environment that utilizes AWS services such as EMR and Quicksight. People Tech Group helped develop a solution that delivered fast, accurate results that users could visualize quickly to perform projections and analysis.
About the Customer:
Elliptica provides artificial intelligence products to help organizations grow with next-generation AI products. Elliptica uses Machine Learning, Deep Learning, and NLP (Natural Language Processing) tools and algorithms to help companies develop solutions. The team has profound knowledge and experience designing, implementing, and integrating Artificial Intelligence products within the customer’s business environment.
About the Partner:
People Tech Group is a full-scale Advanced Tier Consulting Partner with proven expertise in DevOps projects, cloud application development, and infrastructure migrations. With Elliptica, People Tech Group has developed modern Machine Learning and Deep Learning capabilities that give it a unique position.
Why AWS:
AWS provided People Tech Group and Elliptica the ability to build Prophecy and create an easy to deploy, easy to use forecasting solution. AWS also provided us with the support that Prophecy needs to understand better and create a resilient architecture.
Why the Customer Chose the Partner:
People Tech Group has a proven track record of creating and implementing a successful solution with AWS. Elliptica chose to work with People Tech because of their longstanding relationship with AWS, their expertise in AWS services, and their knowledge of time-series forecasting.
Engagement Scope
Elliptica’s goal was to create Prophecy, a new forecasting solution that addresses the following challenges for their customers:
Solution Requirements:
- Prophecy must be able to process upwards of 100g of data in one job run
- Training of Machine Learning algorithms can take no longer than 24 hours
- Prophecy job must follow all AWS standard practices for security
- Availability and Disaster Recovery
- Monitor training jobs and support be readily available
- Costs must be less than 3 cents per 10,000 records
- Visualizations of forecasts must be readily available
Proposed Solution
People Tech Group and Elliptica leveraged AWS to build Prophecy and create an easy to deploy, easy to use forecasting solution. AWS also provided us with the support that Prophecy needs to understand better and create a resilient architecture Prophecy’s solution leveraged quick deployment of the following resources using AWS CDK:
- Step Functions for monitoring and job orchestration
- S3 for initial data storage and data staging
- AWS Glue for ETL and data merging, as well as serverless, scalable compute
- EMR provided easy to configure clusters to perform AutoML on a sample of the submitted data, allowing Prophecy to select the best algorithm for the given data quickly
- Provision of an EMR cluster to be optimized for the chosen ML algorithm
- Predictions, once complete were, saved back to S3 for quick visualization in QuickSight
Impact
Prophecy incorporates a user-friendly interface with state-of-the-art machine learning. It is simple enough to use while offering faster, more accurate results than its competitors. With one click, customers can deploy the required resources to their AWS account and set up needed permissions. Prophecy offers end-to-end automation: Submit your data and Prophecy will determine the best-suited algorithm, create optimized clusters and deliver results quickly. Cluster scaling is handled automatically, with no action required on the part of the user. Once completed, your forecasted data will be available in several file formats, allowing users to move from predictions to business
analysis quickly.
Objective
Global Retailer Challenge
A global grocery retailer wanted to reduce food wastage across all of its stores
- Global grocery retailer that wants to modernize their supply chain to help with over/under stocking
of perishable goods - By reducing waste, they can cut costs and perform more sustainable business
Outdated forecasting methods using statistical calculations have led to increased food wastage in volatile times
They are looking to modernize by using AI-driven forecasting, allowing for more accurate demand projections
that reduce waste and increase profitability
General pitfalls for current forecasting techniques in Retail
- Relying on intuition and manual processes
- Poor Forecasts using only statistical methods
- Not accounting for promotional events may increase the error margin of the forecasts
- Does not account for customer behavior
- Inability to obtain real-time insights
Solution
Traditional methods aimed to solve the loss of sales due to understocking and reduced profits due to overstocking; they do not fully consider changing customer demands and product volatility.
By leveraging available data about products, brands, and customers, Prophecy, with it AI/ML framework for accurate forecasting, helped reduce over/under stocking, not just in grocery retail but in all retail forecasting
Takeaways
- Prophecy increased the accuracy of time-series forecasting by up to 30% compared to cloud forecasting competitors
- Allowed for easy connectivity with a company’s existing AWS infrastructure
- By using built-in AWS fault tolerance, can quickly restart jobs in case of disaster
- Easily compare forecast predictions with incoming results using QuickSight
To know more about People Tech’s Prophecy solution, visit http://peopletech.com/prophecy/
