Data and Python source code for replicating analyses of "Measurement and comparison of distributional shift with applications to ecology, economics, and image analysis"
The concentration of a discrete or continuous distribution of frequencies or probabilities toward a lower bound is a conceptually simple property that historically lacks a global descriptive statistic. We recently termed this property ‘shift’ and defined it as the distance of a central tendency from an upper bound, expressed as a proportion of a finite range. The source code of this repository pertains to our recent study of the shift statistics that we develop.
Fig1_ShiftSkew.ipynb
This Jupyter notebook file is used to generate analyses of the relationship between shift and skewness (Figure 1 of our manuscript).
Fig2_DistMetics.ipynb
This Jupyter notebook file is used to generate comparative analyses between |Δ𝒮| and established distance metrics (Figure 2 of our manuscript).
Fig3_Fig4_Rarity.ipynb
This Jupyter notebook file is used to generate comparative analyses between 𝒮 and established measures of rarity (Figures 3 and 4 of our study). The analyses use combinatorical feasible sets of integer partitions.
Fig5_Poverty.ipynb
This Jupyter notebook file is used to generate comparative analyses between 𝒮 and established measures of poverty (Figure 5 of our study).
Fig6_ImageAnalysis.ipynb
This Jupyter notebook file is used to generate a relatively basic analysis of a synthetic video using our shift statistics (𝒮, Δ𝒮, |Δ𝒮|) and Wasserstein Distance (Figure 6 of our study).
license
This MIT license applies only to source code. It does not pertain to or exert any rights over, e.g., data, images, or video files.
data
This directory contains a single directory (time_lapse_video), which holds a single file (synthetic_video_no_noise_no_bars.mp4). This video file is used in the `Fig6_ImageAnalysis.ipynb` file.
Final_Figs
This directory contains additional directories that are used to store pdf files of figures produced by Jupyter notebooks.