back to list

Project: Interactive Visual Analytics for Outlier Detection and Data Quality Improvement in Large-Scale Industrial Time-Series

Description

Internship Title
Interactive Visual Analytics for Outlier Detection and Data Quality Improvement in Large-Scale Industrial Time-Series Data

Introduction
ASML is developing machine learning models that learn from thousands of sensor signals collected across its worldwide machine fleet. The quality of these models strongly depends on the quality of the underlying data. This project focuses on identifying and understanding anomalous data points and data quality issues in large-scale industrial time-series data.

Objective
Develop methods and an interactive visual analytics framework for detecting, analyzing, and filtering outliers in high-dimensional time-series datasets, improving the robustness and performance of downstream machine learning models.

Background and Scope
Industrial sensor data are often affected by noise, missing values, sensor failures, and other anomalies that can negatively impact model training and prediction accuracy. The student will investigate state-of-the-art anomaly detection techniques, evaluate their effectiveness on real-world industrial datasets, and design visual analytics solutions that allow engineers and data scientists to explore, validate, and refine detected outliers. The work will combine automated detection algorithms with human-in-the-loop analysis to support informed data cleaning decisions.

Deliverables

·         Literature review of anomaly detection techniques for multivariate time-series data.

·         Evaluation of multiple outlier detection approaches on industrial datasets.

·         Interactive visual analytics prototype for anomaly exploration and validation.

·         Reproducible data-cleaning pipeline for generating improved training datasets.

·         Assessment of the impact of data cleaning on downstream machine learning performance.

·         Final report, thesis, and presentation summarizing methodology and results.

Details

Student

LR

Lauria Rubega

Supervisor

Stef van den Elzen

Secondary supervisor

Nicola Pezzotti

External location

ASML