4 ms·

Visualizing distributions of data

107 points by stared 12y ago

stared 12y ago
I find it nicer than bare-bone matplotlib for plotting (one line of code per standard plot + very pleasant graphically). I knew that it exists for some time (and I had it already installed) but only today I rediscovered it as a really useful tool for default plots in my data analysis / data exploration.
chrishenn 12y ago
The violin plots are cool and look like a great alternative to box plots, I had not seen those before. Sweet!
- stared 12y ago
  It is nice as it shows distributions, not only - a few statistics. However, some shapes of violin plots can be... distracting. ;)
achompas 12y ago
This may be in the notebook or on Seaborn's homepage (http://stanford.edu/~mwaskom/software/seaborn/ http://stanford.edu/~mwaskom/software/seaborn/), but someone compared seaborn to pandas in that it provides a terse API to a great base package (matplotlib and numpy, respectively). I'm looking forward to using this more.
- Thrymr 12y ago
  That's the first time I've ever heard the pandas API described as "terse".
  - achompas 12y ago
    Relative to creating your own ndarray, adding a time index, visualizing the data, etc?
    - Fede_V 12y ago
      Pandas is amazing and I use it all the time, but the API for things like hierarchical multi level indexing is very complex (it has to be - it's a very difficult thing).
mwexler 12y ago
Very cool. I suspect many of these port to R as well, and it would be interesting to compare the code and approach.
skeletonjelly 12y ago
Really loving these ipython visualisation articles! There was one about detecting bubbles in liquid for the purposes of reducing spilling that was very interesting http://nbviewer.ipython.org/github/soft-matter/trackpy-examples/blob/master/notebooks/custom_feature_detection.ipynb http://nbviewer.ipython.org/github/soft-matter/trackpy-examp...
- niels_olson 12y ago
  Yo, that's not an article that's the notebook. How awesome is that? This is for science is going.
- Xcelerate 12y ago
  Wow, that's very impressive! This kind of thing could really accelerate my research if I can get in the groove of it. I just downloaded IJulia (IPython for Julia) and having been playing around with it.
tlarkworthy 12y ago
scatter matrix is my goto high dimensional visualization.

107 points by stared 12y ago

I find it nicer than bare-bone matplotlib for plotting (one line of code per standard plot + very pleasant graphically). I knew that it exists for some time (and I had it already installed) but only today I rediscovered it as a really useful tool for default plots in my data analysis / data exploration.

chrishenn 12y ago

The violin plots are cool and look like a great alternative to box plots, I had not seen those before. Sweet!

stared 12y ago
It is nice as it shows distributions, not only - a few statistics. However, some shapes of violin plots can be... distracting. ;)

achompas 12y ago

This may be in the notebook or on Seaborn's homepage (http://stanford.edu/~mwaskom/software/seaborn/ http://stanford.edu/~mwaskom/software/seaborn/), but someone compared seaborn to pandas in that it provides a terse API to a great base package (matplotlib and numpy, respectively). I'm looking forward to using this more.

Thrymr 12y ago
That's the first time I've ever heard the pandas API described as "terse".
- achompas 12y ago
  Relative to creating your own ndarray, adding a time index, visualizing the data, etc?
  - Fede_V 12y ago
    Pandas is amazing and I use it all the time, but the API for things like hierarchical multi level indexing is very complex (it has to be - it's a very difficult thing).

mwexler 12y ago

Very cool. I suspect many of these port to R as well, and it would be interesting to compare the code and approach.

skeletonjelly 12y ago

Really loving these ipython visualisation articles! There was one about detecting bubbles in liquid for the purposes of reducing spilling that was very interesting http://nbviewer.ipython.org/github/soft-matter/trackpy-examples/blob/master/notebooks/custom_feature_detection.ipynb http://nbviewer.ipython.org/github/soft-matter/trackpy-examp...

niels_olson 12y ago
Yo, that's not an article that's the notebook. How awesome is that? This is for science is going.
Xcelerate 12y ago
Wow, that's very impressive! This kind of thing could really accelerate my research if I can get in the groove of it. I just downloaded IJulia (IPython for Julia) and having been playing around with it.

tlarkworthy 12y ago

scatter matrix is my goto high dimensional visualization.