[FIXED] Python programming and group by

Issue

I have very long table, here is an example:

and what do I want is to plot according to image number. given as 1,2,3 and scatter x y (given in column 2 and 3) coordinates on the plot. and color should be according to Type like Green for P and red for F and so on.

Plus I want them to plot separately not in the graph and save them like image.1 image2 and image 3.

Please suggest a possible solution.

I did grouping and now I'm stuck here.

I did until now is following:

for index, group in data.groupby(['image number']):
    
    group.plot(x='x',y='y', label=index, kind='scatter')

Please help me to add colors in the plot according to MP column.

Solution

So from what I understand, your dataframe looks something like this:

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

df = pd.DataFrame([
    {"number": 1, "x": 2391, "y": 1994, "MP": "PP"},
    {"number": 1, "x": 530, "y": 2707, "MP": "PP"},
    {"number": 1, "x": 467, "y": 3208, "MP": "PP"},
    {"number": 2, "x": 952, "y": 656, "MP": "PE"},
    {"number": 2, "x": 3678, "y": 1409, "MP": "PE"},
    {"number": 3, "x": 2950, "y": 177, "MP": "PMMA"},
    {"number": 3, "x": 1039, "y": 306, "MP": "f"},
    {"number": 3, "x": 2726, "y": 1827, "MP": "PMMA"},
    {"number": 3, "x": 3578, "y": 2129, "MP": "PMMA"},
    {"number": 3, "x": 3708, "y": 3624, "MP": "f"},
    # rest of dataframe entries...
    ])

If you want to make a scatter plot of these using matplotlib, you can just iterate over the dataframe and use plt.scatter to plot individual particles like so:

scheme = {
    "PP": "m", # PP is magenta
    "PE": "g", # PE is green
    "PMMA": "c", # PMMA is cyan
    "f": "r", # f is red
    # put more colours here...
    }

def plot_single(df, num, scheme):
    for i in range(len(df.index)):
        if df["number"][i] == num:
            plt.scatter(x=df["x"][i], y=df["y"][i], c = scheme[df["MP"][i]] if df["MP"][i] in scheme else "k")
    # V change these as needed V
    plt.title("Particle "+str(num))
    plt.xlabel("x")
    plt.ylabel("y")
    # ^ change these as needed ^
    plt.show()

As an example, if you try plot_single(df, 1, scheme) with the small snippet of the dataframe, you will get this graph:

To further customize your scatter plot, you can add a legend, change the colour scheme, tinker with the scatter plot, etc.

EDIT:

if you want all the scatter plots on one graph, you can do it with a function like so:

def plot_all(df, scheme):
    for i in range(len(df.index)):
        plt.scatter(x=df["x"][i], y=df["y"][i], c = scheme[df["MP"][i]] if df["MP"][i] in scheme else "k")
    plt.title("All Particles")
    plt.xlabel("x")
    plt.ylabel("y")
    plt.show()

And you can call it by typing plot_all(df, scheme) into your shell.

Answered By - Lucas Roy