A Visual Analytics Framework for Reviewing Multivariate Time-Series Data with Dimensionality Reduction

Demonstration Videos, Source Code, and Datasets

1. Paper

Our paper is available on [arXiv].


2. Demo Video

Video 1: A demonstration of MulTiDR visual interface.


3. Source Code

The source code for the back-end algorithms and web-based visual interface is available from [github].


4. Datasets

1. US Air Quality [1] (Sect. 4 and Case Study 1) [zip file] [original source]

2. MHEALTH: Physical Acitivity Measures [2, 3] (Case Study 2) [zip file], [original source]

3. High School Dynamic Contact Networks [4] (Case Study 3) [zip file], [original source]

* If you use these datasets for your publication, please follow the citation policy of each original source.
* 3rd-order tensor obeject data, tensor.npy file, can be loaded with NumPy: numpy.load(FILE_PATH)

References

[1]

US Environmental Protection Agency. Air Quality System Data Mart [internet database] available via https://www.epa.gov/airdata. Accessed April 28th, 2020.

[2]

Banos et al., mHealthDroid: A Novel Framework for Agile Development of Mobile Health Applications. Proc. IWAAL, 2014.

[3]

Banos et al., I. Design, Implementation and Validation of a Novel Open Framework for Agile Development of Mobile Health Applications. BioMedical Engineering OnLine, vol. 14, no. S2:S6, pp. 1-20, 2015.

[4]

Fournet and Barrat, Contact Patterns among High School Students, PLoS ONE 9(9):e107878, 2014.


5. Parameters Used for Case Studies

See sample.py in github repository for the running example with the parameters below.

1. Case Study 1 (US Air Quality)

PCA: n_components=1, normalization is applied for the input.

UMAP: n_neighbors=7, min_dist=0.15

2. Case Study 2 (MHEALTH)

PCA: n_components=1, normalization is applied for the input.

UMAP: n_neighbors=7, min_dist=0.15

3. Case Study 3 (High School Network)

PCA: n_components=1, normalization is applied for the input.

UMAP: n_neighbors=15, min_dist=0.0

4. Case Study 4 (K Computer Log)

PCA: n_components=1, normalization is applied for the input.

UMAP: n_neighbors=5, min_dist=0.5