Data Analytics, Software Development

Error Reductions Give Better Cash Flow Predictions

Predictive model determines the amount of precious metal in variety of catalytic converters to aid recycling and optimise sales revenue.

About

Founded in 1887, FJ Church & Sons Ltd is an international business dealing in scrap trading who specialise in Non-Ferrous Catalytic Converters and Weee materials. Working from their head office in Rainham, Essex, they were one of the first British scrap merchants to trade internationally.

Having worked with Objective IT for a number of years with developing our online buying portal – the next step for us to improve our customer experience was to integrate an app to the service. This would allow for faster searching , a simpler display and better functionality for our customers using the app ‘on the road’
Dafydd Dylan, FJ Church & Sons Ltd

The Challenge

Recycling catalytic converters for precious metals and materials is an important process at FJ Church. However, the previous mapping method used to predict the amount of metal in each converter used a rule of thumb and therefore was not hugely accurate.

The Solution

Regression analysis was performed by Objective for FJ Church to help reduce the error in the process of removing precious metals from catalytic converters. This resulted in a model being build by the team, improving the accuracy of predicting the content of each converter. This has moved the team from rule of thumb to an intelligent, data driven solution.

The Results & Benefits

  • The new model resulted in a 22% reduction in the mean average error, meaning the team could get a better prediction of the precious metals available when recycling each catalytic converter.
  • This has led to FJ Church gaining a better understanding of future cash flows.

The Technology

R

Python

Other relevant content you maybe interested in