Case study
Wheel Defects Caught Before the Scrap Bin for a Turkish Manufacturer
How SquareShift built real-time dashboards and machine learning models that forecast faulty wheels earlier, helping a Turkish wheel manufacturer catch quality problems before they reach the scrap bin.
Book a session →
Client
A leading wheel manufacturer in Turkey that supplies some of the world's largest automobile manufacturers.
Industry
Manufacturing
Services
Stack
AWS · Machine Learning · Prescriptive Analytics · Apache Spark · PySpark · Tableau · Amazon S3 · AWS Glue
How SquareShift delivered it.
The challenge
About 8% of the wheels the manufacturer produced were failing quality checks and being scrapped. Quality control alone wasn't catching the problem early enough — it counted the damage after the fact instead of catching it while the line was still running.
What we delivered
SquareShift built real-time dashboards from thousands of process variables to track line performance and flag process drift as it happens, instead of only reviewing quality after a wheel was already scrapped.
On top of that, the team built machine learning models that forecast faulty wheels earlier and suggest which process variable to adjust, giving line operators a specific action instead of a general alert.
The outcome
The manufacturer now sees process drift and forecasted defects while the line is still running, instead of only after wheels are scrapped.
Catching a defect after the wheel is scrapped is quality control. Catching it while the process variable is still drifting is the difference prescriptive analytics makes.
AI & Machine Learning Practice Lead, SquareShift
Related reading: building plan review automation with AI and vehicle damage assessment from photos.
