Demand Forecasting for the chemical industry

Chemical
Daten wissen case study image

Business Impact

33% revenue growth

10% error rate reduced.


Customer Facts

Location: Mumbai India

Industry: Chemicals

Daten wissen case study image

Problem Statement

Problems we Acquired

The supply chain team was using traditional methods of forecasting. They were generating a huge error in their forecasting and production planning hence they wanted a robust forecasting solution.

Challenges

Obstacles we faced

  • High number of SKUs
  • Some external factors weren’t captured by the team

Technologies

Tech stack we used

Technology use

PyTorch

Technology use

Django

Technology use

MySQL

Technology use

Python

Technology use

Qlik

Solution

How we made end solution

  • Existing data from SAP and Tally of past 10 years of sales were taken
  • Data engineering and data cleaning, missing data removal
  • Identification of internal and external factors affecting sales
  • Designing of Deep Learning model for forecasting using combination of LSTM and other algorithms
  • Creation and connection of dashboard with SAP to get forecasting in real-time

Result

Outcome we get

  • Customized dashboard to see the visual data
  • Customer was able to predict the product demand accurately