Make a Dynamic Heat Map

📊 Plotly Python
1

I want to make a dynamic heat map which changes with time. I use Python. Can I use the Plotly lib to achieve this?

2

@YellowClothes You can animate a heatmap using Python Plotly. Here is an example: https://plot.ly/~empet/14138 Click the Play button to activate it…

3

What about animating a series of heatmaps objects from package plotly.graph_objs.Heatmap ? is that feasible with plotly ?

4

@nirre Do you mean animating subplots of heatmaps?

5

I have heatmap plots generated with pltly.graph_objs as in this code snipet:
from plotly.offline import plot
import plotly.graph_objs as go

dirpath = os.path.join(os.getcwd(), “weights_visual_raw_data”)
files = [s for s in os.listdir(dirpath) if os.path.isfile(os.path.join(dirpath, s))]

for i, epoch in enumerate(files):
pkl_file = open(os.path.join(os.getcwd(), “weights_visual_raw_data”, epoch), ‘rb’)
weights = pickle.load(pkl_file)
pkl_file.close()
weights = [x for x in weights if x]
for l, layer in enumerate(weights):
if l >= len(weights)-2:
layer_weights = layer[0]
else:
layer_weights = layer[1]
trace = go.Heatmap(z=layer_weights)
data = [trace]
plot(data, auto_open=False, filename=os.path.join(os.getcwd(),
“weights_dynamic_visualization”,
“layer”+str(l),
“weights_epoch_” + str(i)))
In the end I need to play these plots in animation sequence (maybe with a slider that can iterate between different heamaps)

6

Yes, you can define the list of frames consisting in data that you plot successively, in the code above.

7

Could you please demonstrate with a code snippet ? thanks

8

Also, I’m struggling to understand is whether I need to define columns and a grid for my heatmaps plot or can I just assign each frame with the plot as in
trace = go.Heatmap(z=layer_weights)
data = [trace]
plot(data, auto_open=False, filename=’./’).

In case I need to define columns and grid, how to I define a column for a plotly.graph_objs heatmap object ?