Summer-Project

merge Interactive Visual Analytics tools with Data Science tools, prepare for MS thesis Platform Requirements and Packages

1. Python 3
2. Jupyter Notebook
3. Python packages: ipywidgets, matplotlib, numpy, pandas, sklearn, random

Code Design

1. Initialization of sliders
2. MDS function

• Convert high-d data to 2D data

3. Interactive Plot with Sliders

• Allow users to drag the sliders to change the weights
• Then the scatter plot will be changed simultaneously

4. Draggable Plot with Updated Weights

• The input weights in this section is the updated version from the 3th step
• Allow the users to drag any points on the plot to any place

5. Retrieve the Old and New Coordinates

• Retrieve the coordinates of points before and after being dragged by users

6. Inverse lowD to highD

• Proposal：produce a new weight by moving to a random direction with a specific step
• Stress: calculate the stress between old low-d data and new low-d data calculated by high-d data
• invMDs: produce a weight list from the inverse process of MDS, which creates a new high dimensional data.