Bin and count data to normalize Y axis - python-3.x

Question

I have this area and diameter data:
DIAMETER AREA
0 3.039085 1230000
1 2.763617 1230000
2 2.052176 1230000
3 9.498093 1230000
4 2.680360 1230000
I want to bin the data by 1 (2-3,3-4 etc.) and count the number of diameters that are in those bins so that it's organized something like this:
2-3 3-4 4-5 5-6 6-7 7-8 8-9 9-10
3 1 0 0 0 0 0 1
My end goal is to then grab these counts and divide them by the area in order to normalize the counts.
Lastly I will plot the normalized counts (y) by the bins (x).
I tried to use a method using pd.cut but it didn't work

Answer 1

To do this you are wanting a histogram. This can easily be achieved with the hist method of your pandas DataFrame (this itself, just uses the hist plot method from Matplotlib, which uses NumPy's histogram function), for example:
import pandas as pd
from matplotlib import pyplot as plt
# create a version of the data for the example
data = pd.DataFrame({"DIAMETER": [3.039085, 2.763617, 2.052176, 9.498093, 2.680360]})
fig, ax = plt.subplots() # create a figure to plot onto
bins = [2, 3, 4, 5, 6, 7, 8, 9, 10] # set the histogram bin edges
data.hist(
column="DIAMETER", # the column to histogram
bins=bins, # set the bin edges
density=True, # this normalises the histogram
ax=ax, # the Matplotlib axis onto which to plot
)
fig.show() # show the plot
This gives:
where the Pandas hist function will automatically add a plot title based on the column name.
If you don't specify the bins keyword, then it will automatically generate 10 bins bounded by the range of your data, but these will not necessarily be integer spaced. If you wanted to ensure integer spaced bins for arbitrary data, then you could set the bins with:
import numpy as np
bins = np.arange(
np.floor(data["DIAMETER"].min()),
np.ceil(data["DIAMETER"].max()) + 1,
)
If you want a plot like suggested in your other post, then purely with NumPy and Matplotlib you could do:
import numpy as np
from matplotlib import pyplot as plt
# set bins
bins = np.arange(
np.floor(data["DIAMETER"].min()) - 1, # subtract 1, so we have a zero count bin at the start
np.ceil(data["DIAMETER"].max()) + 2, # add 2, so we have a zero count bin at the end
)
# generate histogram
counts, _ = np.histogram(
data["DIAMETER"],
bins=bins,
density=True, # normalise histogram
)
dx_2 = 0.5 * (bins[1] - bins[0]) # bin half spacing (we know this is 1/2, but lets calculate it in case you change it!)
# plot
fig, ax = plt.subplots()
ax.plot(
bins[:-1] + dx_2,
counts,
marker="s", # square marker like the plot from your other question
)
ax.set_xlabel("Diameter")
ax.set_ylabel("Probability density")
fig.show()
which gives: