imagesc

• imagesc is an Python package to create heatmaps. Various methods to create a heatmap are implemented, each with specific properties that can help to easily create your heatmap. The fast and fastclean method is optimized for speed, the cluster method provides clustering, the seaborn method contains many configuration settings, and finally, the plot as good as possible the imagesc from matlab.

Functions in imagesc

# data is your numpy array
fig = imagesc.seaborn(data)
fig = imagesc.cluster(data)
fig = imagesc.fast(data)
fig = imagesc.clean(data)
fig = imagesc.plot(data)
status = imagesc.savefig(fig)

Contents

• Installation
• Requirements
• Quick Start
• Contribute
• Citation
• Maintainers
• License

Installation

• Install imagesc from PyPI (recommended). imagesc is compatible with Python 3.6+ and runs on Linux, MacOS X and Windows.
• It is distributed under the MIT license.

Requirements

pip install numpy pandas matplotlib seaborn
Note that: seaborn is only required when using **seaborn** or **cluster** functions.
or
pip install -r requirements.txt

Quick Start

pip install imagesc

• Alternatively, install imagesc from the GitHub source:

git clone https://github.com/erdogant/imagesc.git
cd imagesc
python setup.py install

Import imagesc package

import imagesc as imagesc

seaborn - The heatmap implemented using seaborn contains a large number of configurations possibilities.

• Slow when using large datasets.
• Grid is aligned to the cells
• See here for all parameters: https://seaborn.pydata.org/generated/seaborn.heatmap.html

df = pd.DataFrame(np.random.randint(0,100,size=(10,20)))
A = imagesc.seaborn(df.values, df.index.values, df.columns.values)
B = imagesc.seaborn(df.values, df.index.values, df.columns.values, annot=True, annot_kws={"size": 12})
C = imagesc.seaborn(df.values, df.index.values, df.columns.values, annot=True, annot_kws={"size": 12}, cmap='rainbow')
D = imagesc.seaborn(df.values, df.index.values, df.columns.values, annot=True, annot_kws={"size": 12}, cmap='rainbow', linecolor='#ffffff')

A B C D

cluster - The heatmap created using the cluster implementation is usefull when you desire to cluster your data.

• Default distance setting: metric="euclidean", linkage="ward"
• Slow for large data sets
• Grid is aligned to the cells
• Possibilities to tweak
• Possible arguments: https://seaborn.pydata.org/generated/seaborn.clustermap.html

df = pd.DataFrame(np.random.randint(0,100,size=(10,20)))
fig_C1 = imagesc.cluster(df.values, df.index.values, df.columns.values)
fig_C2 = imagesc.cluster(df.values, df.index.values, df.columns.values, cmap='rainbow')
fig_C3 = imagesc.cluster(df.values, df.index.values, df.columns.values, cmap='rainbow', linecolor='#ffffff')
fig_C4 = imagesc.cluster(df.values, df.index.values, df.columns.values, cmap='rainbow', linecolor='#ffffff', linewidth=0)
imagesc.savefig(fig_C1, './docs/figs/cluster4.png')

C1 C2 C3 C4

fast - The heatmap created using the fast implementation

• Fast
• Not so much tweakable
• Grid is not aligned to the cells
• Possible arguments: https://matplotlib.org/3.1.1/api/_as_gen/matplotlib.axes.Axes.pcolorfast.html

df = pd.DataFrame(np.random.randint(0,100,size=(10,20)))
fig_F1 = imagesc.fast(df.values, df.index.values, df.columns.values)
fig_F2 = imagesc.fast(df.values, df.index.values, df.columns.values, grid=False)
fig_F3 = imagesc.fast(df.values, df.index.values, df.columns.values, grid=False, cbar=False)
fig_F4 = imagesc.fast(df.values, df.index.values, df.columns.values, grid=True, cbar=False)
fig_F5 = imagesc.fast(df.values, df.index.values, df.columns.values, cmap='rainbow')
fig_F6 = imagesc.fast(df.values, df.index.values, df.columns.values, cmap='rainbow', linewidth=0.5, grid=True)
imagesc.savefig(fig_C1, './docs/figs/fast1.png')

F1 F2 F3 F4 F5 F6

fastclean - The heatmap created using the fastclean implementation is fast and clean with almonst no configurations settings.

• Fast
• Tweakable not so much
• No Grid
• Possible arguments: https://matplotlib.org/3.1.1/api/_as_gen/matplotlib.axes.Axes.pcolorfast.html

df = pd.DataFrame(np.random.randint(0,100,size=(10,20)))
fig_FC1 = imagesc.fastclean(df.values)
fig_FC2 = imagesc.fastclean(df.values, cmap='rainbow')
imagesc.savefig(fig_C1, './docs/figs/fastclean1.png')

F1 F2

plot - The heatmap created using the plot implementation will behave more-or-less as the one of matlab.

• Medium speed
• Tweakable but less then seaborn
• Grid is aligned to the cells
• Possible arguments: https://matplotlib.org/3.1.1/api/_as_gen/matplotlib.pyplot.imshow.html

df = pd.DataFrame(np.random.randint(0,100,size=(10,20)))
fig_M1 = imagesc.plot(df.values)
fig_M2 = imagesc.plot(df.values, cbar=False)
fig_M3 = imagesc.plot(df.values, cbar=False, axis=False)
fig_M4 = imagesc.plot(df.values, cbar=False, axis=True, linewidth=0.2)
fig_M5 = imagesc.plot(df.values, df.index.values, df.columns.values)
fig_M6 = imagesc.plot(df.values, df.index.values, df.columns.values, cbar=False, linewidth=0.2)
fig_M7 = imagesc.plot(df.values, df.index.values, df.columns.values, grid=True, cbar=False, linewidth=0.2)
fig_M8 = imagesc.plot(df.values, df.index.values, df.columns.values, grid=False, cbar=False, linewidth=0.2)
fig_M9 = imagesc.plot(df.values, df.index.values, df.columns.values, grid=True, cbar=False, linewidth=0.8, linecolor='#ffffff')
fig_M10 = imagesc.plot(df.values, df.index.values, df.columns.values, grid=True, cbar=False, linewidth=0.8, linecolor='#ffffff', cmap='rainbow')
imagesc.savefig(fig, './docs/figs/plot10.png')imagesc.savefig(fig_C1, './docs/figs/fast1.png')

M1 M2 M3 M4 M5 M6 M7 M8 M9 M10

Speed:

The heatmap based on plot will behave more-or-less as the one of matlab.

import matplotlib.image as mpimg
img=mpimg.imread('./docs/figs/lenna.png')

fig = imagesc.fastclean(img)
# runtime 1.49

fig = imagesc.fast(img, cbar=False, axis=False)
# runtime: 2.931 seconds

fig = imagesc.plot(img, linewidth=0, cbar=False)
# runtime: 11.042

**fast** **fastclean** **plot**

Citation

Please cite imagesc in your publications if this is useful for your research. Here is an example BibTeX entry:

@misc{erdogant2019imagesc,
  title={imagesc},
  author={Erdogan Taskesen},
  year={2019},
  howpublished={\url{https://github.com/erdogant/imagesc}},
}

References

• seaborn https://seaborn.pydata.org/generated/seaborn.heatmap.html
• clustermap https://seaborn.pydata.org/generated/seaborn.clustermap.html
• fast and fastclean https://matplotlib.org/3.1.1/api/_as_gen/matplotlib.axes.Axes.pcolor.html
• plot https://matplotlib.org/3.1.1/api/_as_gen/matplotlib.pyplot.imshow.html
• Other https://matplotlib.org/3.1.1/gallery/images_contours_and_fields/image_annotated_heatmap.html
• Colormap https://matplotlib.org/examples/color/colormaps_reference.html

Maintainers

• Erdogan Taskesen, github: erdogant

Contribute

• Contributions are welcome.

© Copyright

See LICENSE for details.