Skip to content

Latest commit

 

History

History
25 lines (20 loc) · 1.29 KB

File metadata and controls

25 lines (20 loc) · 1.29 KB

INTERACTIVE DATA VISUALIZATION OF ELECTRONIC NOSE (Python: Bokeh library)

Dataset consists of signals recorded by R. Huerta et al., [1] via electronic nose consisting of 8 MOX gas sensors and a temperature and a humidity sensor. The electronic nose was placed in a home environment and was exposed to two different stimuli: banana and wine. Thr background activity was also recorded. [1, 2]

The interactive visualization in Bokeh allows for comparison of signals resulting from different stimuli as well as for visualization of these signals decorellated from temperature and humidity.

This project consist of:

  1. Two jupyter notebooks:
    • Data preparation
    • Interactive visualization in Bokeh
  2. Original data files:
    • Dataset_Split10min.npy
    • Dataset_Split10min_hashtable.npy
    • HT_Sensor_metadata.dat
  3. Modified data files (code in notebooks):
    • dataset_pd.pkl
    • dataset_pd_decorr.pkl
    • metadata_sub.pkl

References:

  1. R. Huerta et al., Chemom. Intell. Lab. Syst. 157, 169-176 (2016).
  2. UCI Machine Learning Repository: https://archive.ics.uci.edu/ml/datasets/Gas+sensors+for+home+activity+monitoring

Alt Text