I am using Tensorboard 2 to visualize my training data and I am able to save scalar plots to disk. However, I am unable to find a way to do this for histogram plots (tf.summary.histogram).
Is it possible to save histogram plots from Tensorboard 2 to disk, just like it is possible to do with scalars? I have looked through the documentation and it seems like this is not supported, but I wanted to confirm with the community before giving up. Any help or suggestions would be greatly appreciated.
There is an open issue to add a download button for histograms. However, this issue is open for more than 4 years, so I doubt it is getting resolved soon.
A workaround is to use the url that tensorboard would use to get the data.
A short example:
# writing some data to tensorboard
from torch.utils.tensorboard import SummaryWriter
import numpy as np
writer = SummaryWriter('./tmp')
writer.add_histogram('hist', np.arange(10), 0)
Open tensorboard in the browser (here localhost:6006):
Get data as JSON using the template
http://<tb-host>/data/plugin/histograms/histograms?run=<run-name>&tag=<tag-name>.
Here http://localhost:6006/data/plugin/histograms/histograms/?run=.&tag=hist:
Now you can download the data as JSON.
Quick comparison with matplotlib:
import pandas as pd
import json
import matplotlib.pyplot as plt
with open('histograms.json', 'r') as f:
d = pd.DataFrame(json.load(f)[0][2])
fix, axes = plt.subplots(1, 2, figsize=(10, 3))
axes[0].bar(d[1], d[2])
axes[0].set_title('tb')
axes[1].hist(data)
axes[1].set_title('original')
Related
I'm hoping to get some help here. I'm trying to create some simple bar/line graphs from a csv file, however, it gives me an empty graph until I open this csv file manually in excel and change the data type to numeric. I've tried changing the data type with pd.to_numeric but it still gives an empty graph.
The csv that I'm trying to visualise is web data that I scraped using Beautiful Soup, I used .text method do get rid of all of the HTML tags so maybe it's causing the issue?
Would really appreciate some help. thanks!
Data file: https://dropmefiles.com/AYTUT
import numpy
import matplotlib
from matplotlib import pyplot as plt
import pandas as pd
import csv
my_data = pd.read_csv('my_data.csv')
my_data_n = my_data.apply(pd.to_numeric)
plt.bar(x=my_data_n['Company'], height=my_data_n['Market_Cap'])
plt.show()
Your csv file is corrupt. There are commas at the end of each line. Remove them and your code should work. pd.to_numeric is not required for this sample dataset.
Test code:
from matplotlib import pyplot as plt
import pandas as pd
my_data = pd.read_csv('/mnt/ramdisk/my_data.csv')
fig = plt.bar(x=my_data['Company'], height=my_data['Market_Cap'])
plt.tick_params(axis='x', rotation=90)
plt.title("Title")
plt.tight_layout()
plt.show()
I am using python 3.8 on Windows 10; trying to make a plot with about 700M points in it, sound wave analysis. Here: Interactive large plot with ~20 million sample points and gigabytes of data
Vaex was highly recommended. I am trying to use examples from the Vaex tutorial but the graph does not appear. I could not find a good example on Internet.
import vaex
import numpy as np
df = vaex.example()
df.plot1d(df.x, limits='99.7%');
The Vaex documents don't mention that pyplot.show() should be used to display. Plot1d plots a histogram. How to plot just connected points?
I am pretty sure that the vaex documentation explains that the (now deprecated) method .plot1d(...) is a wrapper around matplotlib plotting routines.
If you would like to create custom plots using the binned data, you can take this approach (I also found it in their docs)
import vaex
import numpy as np
import pylab as plt
# Load example data
df = vaex.example()
# Do the binning yourself
counts = df.count(binby=df.x, shape=64, limits='99.7%')
# Take care of the x-axis
limits = df.limits_percentage(df.x, percentage=99.7)
xvals = np.linspace(limits[0], limits[1], num=64)
# Create your custom plot via matplotlib, plotly or your favorite tool
p.plot(xvals, counts, marker='o', ms=5);
The following is my code, but I can't get the plot to show on my Visual Studio Code even though I am running this on the Python Interactive Window, which should usually show a graph plot after running. The tables are showing just fine. I also do not get a default graph which pops up like it normally should. What am I doing wrong?
import yfinance as yf
import pandas as pd
import matplotlib
matplotlib.use('agg')
import matplotlib.pyplot as plt
import talib
df = pd.read_csv('filename.csv')
df = df.sort_index()
macd, macdsignal, macdhist = talib.MACD(df['Close'], fastperiod=12, slowperiod=26, signalperiod=9)
macd = macd.to_list()
macdsignal = macdsignal.to_list()
macdhist = macdhist.to_list()
df['macd'], df['macdsignal'], df['macdhist'] = macd,macdsignal,macdhist
ax = plt.gca()
print(df.columns)
df.plot(kind='line',x='Date',y='macd', color='blue',ax=ax)
df.plot(kind='line',x='Date',y='macdsignal', color='red', ax=ax)
plt.show()
The csv file has data that looks like this
The issue was with matplotlib.use('agg'), which does not support the show() function. This prevented the graph from being displayed on Visual Studio's Interactive Window. The matplotlib.use('agg') method can, however, be used for saving your graph in a .png format.
According to Matplotlib.org, agg is "the canonical renderer for user interfaces, which uses the Anti-Grain Geometry C++ library to make a raster (pixel) image of the figure". More information can be found at this link here
I am new to JupyterLab trying to learn.
When I try to plot a graph, it works fine on jupyter notebook, but does not show the result on jupyterlab. Can anyone help me with this?
Here are the codes below:
import pandas as pd
import pandas_datareader.data as web
import time
# import matplotlib.pyplot as plt
import datetime as dt
import plotly.graph_objects as go
import numpy as np
from matplotlib import style
# from matplotlib.widgets import EllipseSelector
from alpha_vantage.timeseries import TimeSeries
Here is the code for plotting below:
def candlestick(df):
fig = go.Figure(data = [go.Candlestick(x = df["Date"], open = df["Open"], high = df["High"], low = df["Low"], close = df["Close"])])
fig.show()
JupyterLab Result:
Link to the image (JupyterLab)
JupyterNotebook Result:
Link to the image (Jupyter Notebook)
I have updated both JupyterLab and Notebook to the latest version. I do not know what is causing JupyterLab to stop showing the figure.
Thank you for reading my post. Help would be greatly appreciated.
Note*
I did not include the parts for data reading (Stock OHLC values). It contains the API keys. I am sorry for inconvenience.
Also, this is my second post on stack overflow. If this is not a well-written post, I am sorry. I will try to put more effort if it is possible. Thank you again for help.
TL;DR:
run the following and then restart your jupyter lab
jupyter labextension install #jupyterlab/plotly-extension
Start the lab with:
jupyter lab
Test with the following code:
import plotly.graph_objects as go
from alpha_vantage.timeseries import TimeSeries
def candlestick(df):
fig = go.Figure(data = [go.Candlestick(x = df.index, open = df["1. open"], high = df["2. high"], low = df["3. low"], close = df["4. close"])])
fig.show()
# preferable to save your key as an environment variable....
key = # key here
ts = TimeSeries(key = key, output_format = "pandas")
data_av_hist, meta_data_av_hist = ts.get_daily('AAPL')
candlestick(data_av_hist)
Note: Depending on system and installation of JupyterLab versus bare Jupyter, jlab may work instead of jupyter
Longer explanation:
Since this issue is with plotly and not matplotlib, you do NOT have to use the "inline magic" of:
%matplotlib inline
Each extension has to be installed to the jupyter lab, you can see the list with:
jupyter labextension list
For a more verbose explanation on another extension, please see related issue:
jupyterlab interactive plot
Patrick Collins already gave the correct answer.
However, the current JupyterLab might not be supported by the extension, and for various reasons one might not be able to update the JupyterLab:
ValueError: The extension "#jupyterlab/plotly-extension" does not yet support the current version of JupyterLab.
In this condition a quick workaround would be to save the image and show it again:
from IPython.display import Image
fig.write_image("image.png")
Image(filename='image.png')
To get the write_image() method of Plotly to work, kaleido must be installed:
pip install -U kaleido
This is a full example (originally from Plotly) to test this workaround:
import os
import pandas as pd
import plotly.express as px
from IPython.display import Image
df = pd.DataFrame([
dict(Task="Job A", Start='2009-01-01', Finish='2009-02-28', Resource="Alex"),
dict(Task="Job B", Start='2009-03-05', Finish='2009-04-15', Resource="Alex"),
dict(Task="Job C", Start='2009-02-20', Finish='2009-05-30', Resource="Max")
])
fig = px.timeline(df, x_start="Start", x_end="Finish", y="Resource", color="Resource")
if not os.path.exists("images"):
os.mkdir("images")
fig.write_image("images/fig1.png")
Image(filename='images/fig1.png')
I want to store a matplotlib figure and load it later to use it interactively. To be more specific, I want to be able to use zoom in this figure.
I am using pickle to dump the figure handle into a file.
I then load the figure later using pickle, but the zoom does not work after loading the file even though I can use zoom in the figure before pickling it.
Here's a sample script that illustrates my problem.
import matplotlib.pyplot as plt
import pickle
import numpy as np
import os
import time
# Create Plot Data
x = np.arange(100)
# Create Figure, Axes and plot
fig1,axes1 = plt.subplots()
axes1.plot(x)
# Pickle plot
fileName = os.getcwd() + "/img"\
+ time.asctime(time.localtime()) + ".pickle"
with open(fileName,'wb') as pickle_file:
pickle.dump(fig1,pickle_file)
plt.show() # ZOOM WORKS HERE
plt.close()
# Load pickled plot
with open(fileName,'rb') as read_pickle:
fig_handle = pickle.load(read_pickle)
plt.show() # ZOOM DOES NOT WORK HERE
Zooming into the image before Pickle
Non-Zoomable image after Pickle
Version:
Python 3.7.0
Matplotlib 3.0.0
Pickle 4.0
Is this a limitation with Pickling matplotlib figure? Or is there something I can do to load/dump the figure in a zoomable way?
As #ImportanceOfBeingErnest pointed out I fixed the problem by changing my backends from MacOSX to TkAgg.
Here's what I did.
import matplotlib
matplotlib.get_backend() # Get the current backend
'MacOSX'
# Get Location where configuration file was loaded from.
matplotlib.matplotlib_fname()
'/usr/local/lib/python3.7/site-packages/matplotlib/mpl-data/matplotlibrc'
# Edit -> backend:TkAgg
matplotlib.get_backend()
'TkAgg'