Assignment 4

Code

import matplotlib.pyplot as plt
import matplotlib.ticker as ticker
import numpy as np; np.random.seed(22)
import seaborn as sns; sns.set(color_codes=True)
import pandas as pd
import math

Code

datadir = '/Users/codyappa/Documents/GitHub/Data_Science_Portfolio/posts/MarksChannels/Sheet 3-Table 1'

Code

df = pd.read_csv('Sheet 3-Table 1.csv')

Figure 1

The following figure shows chromosomal counts of untreated and induced samples of Euo. By using these counts we can determine that the overall replication rate of the two treatments is the same.

Marks/Channels

I used points to show the individual timepoints and then lines to connect them and show the correlation between the two. I then used differing colors to indicate the different datasets. Length of the errorbar is to show diversion from the mean.

Code

with plt.style.context('seaborn-white'):
    fig, ax = plt.subplots(ncols=1)
    sns.pointplot(x=df['hpi'], y=df['pyrG copies/ul'], hue=df['Treatment'], data=df, errwidth=2, capsize=.1, linestyles='-', palette=('orange', 'b'), dodge=True, ax=ax)
    with plt.style.context('classic'):
        ax.legend(loc='upper left', fontsize=10)
        ax.set(ylabel='Cell Count')
ax.ticklabel_format(axis="y", style="sci", scilimits=(0,2))
fig.set_size_inches(4, 4)

Code

plt.show()

Figure 2 Marks/Channels

I made a heat map of the datasets which really doesn’t show the correlation or anything. You can see that as we get towards 48hpi they seem to follow the same ish pattern but really you can’t get any correlations from this.

• Expressive/Effective: the data presented is not shown in a way easily perceivable. Just by looking at the data you can’t really interpret anything unless I were to describe in detail what is occuring. There is too much overlap and many channels thrown at you; what does the color intensity mean? why is there a gap in the middle? what does orange vs blue mean?

• Separability: the first dataset was easily separated by color and connecting lines. The second dataset, however, has a lot of overlap which makes it hard to be able to distinguish individual datapoints or any correlation as you look at it. you can just kinda see the colors vaguely mash together in the same areas.

Code

with plt.style.context('seaborn-white'):
    fig, ax = plt.subplots(ncols=1)
    sns.histplot(x=df['hpi'], y=df['pyrG copies/ul'], hue=df['Treatment'], data=df, ax=ax, linewidth=.9)
    ax.set(ylabel='Cell Count')
    ax.ticklabel_format(axis="y", style="sci", scilimits=(0,2))

Code

plt.show()