yellowbrick
yellowbrick copied to clipboard
Animated Feature Visualizers
Without too much work, it appears that we can create an animated image that captures changes to a dataset over time or some series related domain.
We could accomplish this by using the built in matplotib.animation
Here's an example located here: https://matplotlib.org/2.1.2/gallery/animation/dynamic_image2.html
Here's an example API:
from yellowbrick.features import RadViz
from yellowbrick.contrib import AnimateFeatures
X = some_tabular_data_source
time_column = 'timestamp'
oz = AnimateFeatures(RadViz, X, series_column=time_column, outpath='dynamic_images.mp4')
For anyone interested in this one, it might also be worth it to check out animatplot, a Python package for creating animated visualizations (built on top of Matplotlib).
Hello @ndanielsen and @rebeccabilbro ,
I have been working on this issue for past few days and have got some satisfying results. However, I had some questions related to the project. I would be grateful if you could help me through this.
I have attached my results with two Classes (RadViz and Rank1d). Let me know if some changes are to be made.
New folder (2).zip
@naresh-bachwani. Thank you for this message. Yellowbrick has been on hiatus and just now returning. We will respond asap. In the meanwhile i encourage you to open a PR. Our contributing guide can be found here http://www.scikit-yb.org/en/latest/contributing.html
I am not sure about my solution and needed some insights regarding the code. I want to open a PR once I am sure about the solution.
@naresh-bachwani perhaps we can explore this toward the end of the summer together. Also @ndanielsen may be interested in this.
I would like to have this feature in the yb library! It will be good to have @ndanielsen with us.
@naresh-bachwani I am closely paying attention to this issue. I'll circle back on this issue and hopefully contribute a bit when some of the testsuite issues are resolved.
Keep up the good work on this :1st_place_medal: