로고 로고

SITE MAP

REFERENCE AI·Big Data

REFERENCE

Based on over 20 years of experience in building smart factories,
DL Information Technology Co. Ltd. is creating manufacturing big data analysis
and AI application cases in various industrial fields

AI/Big Data Construction Case_ Chungnam mobility company's intelligent raw/subsidiary materials purchase order demand forecast application case

DL Information Technology Co., Ltd. provides intelligent raw/subsidiary materials purchase order demand forecast for Chungnam mobility enterprises through Chungnam AI regional specialized industry

Raw data Analysis environment/Technology Data collection/Loading Visualization

Raw Data

  • Image order form

  • Public institution websites

Demand company product sales data
Image order form due to ordering party's security policy
External variable data of public institution website (25 types including international oil price, diesel price, consumer price, etc.)

Analysis environment / technology

  • Dataset

  • Learning and verification

Monthly sales volume of each product and selected monthly external factor data
Each external factor is used as Future Regressor and Lagged Regressor separately
When learning, Neural Prophet and Prophet compete to select the model with low MAE

Data Collection / Loading

Converting the image order form into text data using OCR and loading it into the database
To analyze product sales data, convert it into an appropriate format and load it into the database
Automatically collecting external variable data (exchange rate, consumer price index, interest rate, international oil price, etc.) that may affect sales through API and crawling and loading it into the database

Visualization

  • Forecast status dashboard screen

  • Demand forecast result screen

Paint Point

  • Manual method

  • Difficult to respond

  • Increase in
    unnecessary
    inventory

  • Ordering raw/subsidiary
    materials,
    quantity error

Manual work process
Manual production planning
Difficulty in responding to frequent order changes
from ordering parties
Increase in unnecessary inventory to respond
to frequent order changes
Increase in long-term unusable inventory due to ordering
of different raw/subsidiary materials, quantity errors, etc.
Introduction effect
  • Increased work efficiency
    Easy drawing search and management through data DB
    Easy expansion for metadata linkage
    Increased data usability through management system construction
    [Previous] 5 times x 2 people x 2.0 hours = 437,000 won/day
    [Changed] 5 times x 1 person x 2.0 hours = 54,000 won/day
    * Annualized savings: 50,531,000 won/year
    Reduced workload > Expected productivity improvement through job transfer
  • Demand forecast
    Preventing errors in work on similar new designs by increasing the usability
    of existing verified drawings
    Forecasting demand through algorithms utilizing Neural Prophet / Prophet and using the result as basis for production planning
    Machine learning and deep learning using sales data, production plan data, safety stock data, and external data (oil prices, exchange rates, etc.)
Client company