Toggle search
Toggle location

Digital analytics

The automotive industry is entering a new age of analytics!

In BearingPoint’s "Big Data in Automotive" study among 120 decision-makers surveyed from automobile manufacturers and suppliers , 94% of respondents attribute great importance to Big Data & Analytics. At the same time, only 7% of companies are fully using Big Data & Analytics. The scope of application covers all business areas.

Why Predictive Analytics? Why now?

Without Big Data Analytics, companies are blind and deaf, wandering out onto the web like a deer on a freeway.

- Geoffrey Moore, American organizational theorist, management consultant and author

The importance and size of data has increased significantly in recent years. One of the biggest challenges in the automotive industry is the intelligent evaluation and use of large amounts of data in the business environment and in vehicles. Advanced Analytics makes it possible to forecast future developments at an early stage and to identify meaningful patterns.

The focus is on processing and unifying data from different business divisions: only a holistic view shows the cross connections between business areas. That was reflected among the BearingPoint study participants, with three out of four of the opinion that current developments should not be missed and that Big Data & Analytics should be introduced.

Predictive analytics is about predicting the future, in contrast to conventional descriptive analysis which concentrates on the comprehensive analysis of historical data. Using state-of-the-art algorithms such as Machine Learning or Artificial Intelligence, complex patterns in huge amounts of data can be recognized in order to make statements about future events. Prescriptive analytics goes one step further, giving direct recommendations for action, such as real-time offering and pricing.

For example, our unique Quality Navigator enables our customers to increase quality along the entire product development process using artificial intelligence and machine learning:

Quality Navigator – A new age of quality analytics

In addition, Predictive Analytics typically faces the following challenges that require expertise in Big Data & Analytics:

  • Data Engineering:
    • Creation of digital connectivity to industry 4.0 and smart devices.
    • Collecting, preparing and storing data in different formats and systems (Hadhoop, Kafka, Splunk).
  • Data Science:
    • Analysis and definition of relevant data and information using modern tools and algorithms (R, Python, AI, Machine Learning, Deep Learning).
    • Definition of recommendations for action in cooperation with industry and process experts.
    • Target-group specific visualization in reporting tools (BOARD, Tableau, Qlik Sense, SAP BI).
  • Data security
  • Data quality and consistency

Why BearingPoint?

BearingPoint is your suitable partner in Predictive Analytics:

  • We have successfully completed many predictive analytics projects in the automotive industry.
  • We combine specific process know-how along the entire automotive value chain with the use of modern data science technology.
  • We cover the entire service range from data acquisition and analysis to business management recommendations.
  • With our BearingPoint Predictive Workbench, all modern data science tools are available.

Below you will find a selection of use cases in which our Predictive Analytics solutions have successfully identified and recovered potentials.

  • 1. Quality Navigator – Reduction of detection-to-correction cycle

    We provide the Eco system to develop a customer-specific early warning system for quality and warranty issues.

    Problem Description: Client Situation & Pain Points

    • No early warning system for technical actions
    • Root cause analysis for quality problems is not conducted
    • No cost and provision forecasts

    Project Results: Our Approach

    The Quality Navigator covers the entire product form quality to lifecycle with the help of the Predictive Analytics Workbench:

    • Predictive analytics models developed by BearingPoint data scientists and partners to reduce costs and improve early warning systems
    • Implementation of Predictive Diagnostics, Predictive Maintenance, Root Cause Analysis and Provision Management
    • Predefined integrated data model for after-sales, warranty, vehicle and quality data

    Customer Benefits: Measurable and Sustainable

    • Shorter detection-to-correction cycle time
    • Fewer warranty cases
    • Improved vehicle quality
    • Lower costs and improved forecasting
    • High usability and intuitive analysis possibilities by using the most modern dashboard and reporting tools
  • 2. Factory Navigator – Optimization of sales forecasting

    We optimize sales forecasting through hierarchical, combined forecasting methods

    Problem Description: Client Situation & Pain Points

    • Demand structures differ greatly between products und regions
    • Lack of quantitative consideration of temporary demand shocks
    • External influences (hybrid promotion, etc.) difficult to quantify

    Project Results: Our Approach

    The Predictive Analytics module of the Factory Navigator was used to improve the sales forecast:

    • Creation of consistent forecast results across all levels of sales planning (region, country, model, derivative, etc.)
    • Optimization of forecast quality by taking into account demand patterns, trends, seasons, shocks, structural breaks, cluster effects, external factors, etc.
    • Quantitative planning of events and new product launches that are difficult to predict

    Customer Benefits: Measurable and Sustainable

    • Sound basis for sales planning decisions
    • Predictability of external effects (hybrid promotion, etc.) in relation to effect strength and effect duration
    • Consistent sales forecasts across all product structures and geographical structures (no contradictory individual forecasts
  • 3. Optimization of production times for EC screwing by prediction of defects

    Power screwdriver analytics based on big data analytics and sample analysis

    Problem Description: Client Situation & Pain Points

    • Errors in EC screwing are only detected at the end of the screwing process
    • Screwdriving data can be recorded online, but there are no reliable error patterns
    • It is desirable to abort the screwing process immediately after an error has occurred

    Project Results: Our Approach

    • Timely detection of error case on the basis of a pattern recognition close to the equipment in the operating parameters of the EC screwdriver by means of FPGA. Appropriate measures (rework or termination) are indexed.
    • Samples were determined by long-term measurement (big data) and analysis of all available EC screwdriver stations
    • The process can be used for other production technologies

    Customer Benefits: Measurable and Sustainable

    • Reduction of the F-time in the event of an error with the possibility of time-neutral rectification
    • Derivation of predictive maintenance of the EC screwdriver in case of indication from the samples of the operating parameters
    • Production-related quality assurance of the screw connections
  • 4. Analysis of battery data in commercial vehicles

    Analysis of battery data in vehicles to create transparency and failure prevention

    Problem Description: Client Situation & Pain Points

    • Commercial vehicles often fail due to deep discharges of the installed batteries
    • High goodwill costs arise due to multiple replacements despite low unit costs
    • There is too little internal transparency regarding reasons for failure

    Project Results: Our Approach

    • Merge different data sources into one analysis database for the project
    • Creating transparency by identifying usage profiles (overnight stays, loading/unloading cycles, deep discharges)
    • Identification of possibilities for optimizing energy management in commercial vehicles
    • Analysis of relationships between temperatures, charging cycles and battery aging

    Customer Benefits: Measurable and Sustainable

    • Improve customer satisfaction by reducing downtime
    • Increased transparency regarding the use of the battery, which is included in the assessment of customer complaints and goodwill cases
    • Long-term cost savings
  • 5. Failure prognosis during spot welding

    Predicting spot welding failures in vehicle production has enormous potential to prevent production downtime

    Problem Description: Client Situation & Pain Points

    • Welding robots from two production lines fail unpredictably and interrupt production
    • The data records contain inconsistencies
    • Proactive maintenance during shift changes will prevent failures in the future

    Project Results: Our Approach

    • The analysis showed serious errors in documentation and data recording
    • Despite data quality deficiencies, one of five Machine Learning algorithms tested can already predict the first failures
    • Concrete recommendations for structured documentation of incidents lead to a significant increase in quality
    • The forecasting quality can be significantly increased by additional data sources

    Customer Benefits: Measurable and Sustainable

    • Avoidance of a malfunction during five robot maintenance operations
    • Process-related improvement possibilities for the documentation of incidents
    • Prediction of failures in the previous shift enables the reduction of overcapacities
  • 6. Sales Navigator – Forecast of the all-time demand

    Prediction of the all-time demand for spare parts in the automotive industry in the aftersales sector

    Problem Description: Client Situation & Pain Points

    • Reduction of spare parts inventories through better forecasting models
    • Low forecast accuracy
    • So far use of simple statistical methods

    Project Results: Our Approach

    • Formation of parts groups with the same demand characteristics. Creation of a matrix with the following fields: high demand, low demand, sporadic, frequent
    • Preparation of data, determination of predecessor and successor models
    • Implementation of forecast models for the respective parts groups
    • Development of a GUI for the input of control parameters for the calculation

    Customer Benefits: Measurable and Sustainable

    • Decision based on solid forecast models
    • Reduce manual effort in predicting parts inventory
    • Reduction of parts inventories due to more accurate forecast results

© 2018 BearingPoint All Rights Reserved