Michael Thane

Machine Learning · Visual Analytics · Research Engineering

Computational methods and interactive systems for complex scientific data.

Portrait of Michael Thane.
01 / About

About

Computer scientist and PhD candidate working at the intersection of machine learning, computer vision, visual analytics, and research software engineering. I develop computational methods and interactive systems for extracting, analysing, visualising, and interpreting complex scientific data.

My work ranges from image-based tracking, segmentation, and feature extraction to representation learning, dimensionality reduction, clustering, interactive visualisation, and research software development. I have several years of interdisciplinary research experience in behavioural neurobiology and previously worked as a Visiting Researcher at Hokkaido University in Sapporo, Japan.

02 / Research

Research & Professional Status

Previous Research Affiliations

German (native) · English (fluent) · Japanese (intermediate).

03 / Expertise

Technical Expertise

Programming & Engineering

  • Python
  • R
  • JavaScript
  • SQL
  • C++
  • Bash
  • Git
  • Docker
  • Linux

Machine Learning & Data Analysis

  • PyTorch
  • scikit-learn
  • NumPy
  • pandas
  • Supervised Learning
  • Unsupervised Learning
  • Representation Learning
  • Clustering
  • Dimensionality Reduction
  • Time-series Analysis
  • Model Evaluation

Computer Vision

  • OpenCV
  • Image Tracking
  • Segmentation
  • Object Detection
  • Motion Analysis
  • Trajectory Processing

Visual Analytics

  • R Shiny
  • D3.js
  • Plotly
  • Coordinated Views
  • Graph Visualisation
  • Interactive Exploration

Research Software

  • Analysis Pipelines
  • Data Preprocessing
  • Research Prototypes
  • Flask
  • Django
04 / Projects

Projects

Interactive research systems and video walkthroughs demonstrating computational methods applied to complex scientific data.

  • Screenshot of CatNetVis.

    CatNetVis

    Interactive visual analytics for categorical high-dimensional data

    An interactive visual analytics system for exploring categorical high-dimensional data through graph-based visualisation. Semantic network representations support the discovery of relations between categorical attributes, helping users identify meaningful structures, clusters, and associations in complex data sets.

    Outcome: EuroVis Short Paper 2023

    • Visual Analytics
    • Graph Visualisation
    • High-Dimensional Data
    • D3.js
  • Screenshot of KumaQ.

    KumaQ

    Spatiotemporal visualisation of wildlife sightings in Hokkaido

    An interactive map-based visualisation system for exploring wildlife sightings across Hokkaido. Combines spatial, temporal, and contextual information to support the analysis of sighting patterns, risk areas, seasonal dynamics, and route-related exposure.

    Outcome: Research prototype for environmental visual analytics

    • Spatiotemporal Data
    • Leaflet
    • Environmental Data
    • Hokkaido
  • RelationExplorer interface.

    RelationExplorer

    Exploratory analysis of high-dimensional behavioural data

    An interactive visual analytics system for exploratory analysis of high-dimensional behavioural data. Coordinated views and relation-based exploration combine type-aware relation measures, clustering, and interactive filtering to help researchers identify meaningful patterns across numerical and categorical attributes.

    Outcome: VMV 2025 research system

    • Visual Analytics
    • Coordinated Views
    • Relation Discovery
    • Behavioural Data
05 / Publications

Publications

Peer-reviewed articles, preprints, and refereed proceedings. Citation counts and the full bibliography are on Google Scholar and ORCID.

Preprints

  1. Sen, E., Königsmann, S., Besharatifar, M., Ciuraszkiewicz, A., Demirci, S., Güler, A. I., Niewalda, T., Schleyer, M., Thane, M., König, C., Thoener, J., & Gerber, B. (2026). Starvation modulates associative short-term memory of Drosophila in a task-dependent manner. bioRxiv.

    bioRxiv preprint · doi.org/10.64898/2026.06.09.729932

Published

  1. Kolms, J., Blum, K. M., Thane, M., Kurczveil, T., & Lehmann, D. J. (2026). Energy Optimized Green Light Assist in Varying Traffic Scenarios Using Reinforcement Learning. In SUMO Conference.

  2. Thane, M., Blum, K. M., & Lehmann, D. J. (2025). Uncovering Relations in High-Dimensional Behavioral Data of Drosophila melanogaster. In VMV — Vision, Modeling and Visualization.

    doi.org/10.2312/vmv.20251234

  3. Bormann, A., Körner, M. B., Dahse, A.-K., Gläß, S., Irmer, J., Lede, V., Alenfelder, J., Lehmann, J., Hall, D. C. N., Thane, M., Schleyer, M., Kostenis, E., Schöneberg, T., Bigl, M., Langenhan, T., Ljaschenko, D., & Scholz, N. (2025). Intron retention of an adhesion GPCR generates 1TM isoforms required for 7TM-GPCR function. Cell Reports, 44(1), 115078.

    doi.org/10.1016/j.celrep.2024.115078

  4. Thane, M., Blum, K. M., & Lehmann, D. J. (2023). CatNetVis: Semantic Visual Exploration of Categorical High-Dimensional Data with Force-Directed Graph Layouts. In EuroVis Short Papers.

    doi.org/10.2312/evs.20231049

  5. Thane, M., Paisios, E., Stöter, T., Krüger, A.-R., Gläß, S., Dahse, A.-K., Scholz, N., Gerber, B., Lehmann, D. J., & Schleyer, M. (2023). High-resolution analysis of individual Drosophila melanogaster larvae uncovers individual variability in locomotion and its neurogenetic modulation. Open Biology, 13(4), 220308.

    doi.org/10.1098/rsob.220308

  6. Thane, M., Viswanathan, V., Meyer, T. C., Paisios, E., & Schleyer, M. (2019). Modulations of microbehaviour by associative memory strength in Drosophila larvae. PLOS ONE, 14(10), e0224154.

    doi.org/10.1371/journal.pone.0224154

Presentations

  • CatNetVis: Semantic Visual Exploration of Categorical High-Dimensional Data. EuroVis 2023, Leipzig, Germany.
  • Uncovering Relations in High-Dimensional Behavioral Data. VMV, 2025.
Contact

Contact

For collaborations or general inquiries, please write by email.