lrkfallon / using-data-aggregation-and-a-scatter-plot-to-analyse-obama-s-approval-rating-from-2008-to-2017.ipynb

Scatter graph, with median and percentile plots, displaying Obama's approval ratings from 2008 to 2017, using data imported from a CSV file.

Using-Data-Aggregation-to-Analyse-Obama-s-Approval-Rating-from-2008-to-2017

Scatter graph, with median and percentile plots, displaying Obama's approval ratings from 2008 to 2017, using data imported from a CSV file. This was formatted for Jupyter Notebook.

Importing the libraries and modules:

import pandas as pd from matplotlib import pyplot as plt

Importing the data from a .csv file:

data = pd.read_csv('obama.csv', parse_dates=['year_month'])

Using the parse_dates functions will instruct pandas to interpret the year's month as

a series time stamps instead of as a string. A series of time stamps is needed to plot a

time series graph.

data.head()

plt.plot(data.year_month, data.approve_percent, 'o', markersize=2, alpha=0.3) plt.show()

To aggregate the data and summarise it, one method that is effective is to use the mean

approval percentage for each month. This can be acheived by using the groupby function:

data.groupby('year_month').mean()

This shows the mean approval percentage adn mean disapproval percentage for each month.

To show the mean:

data_mean = data.groupby('year_month').mean()

To also show the median as well as the mean:

data_median = data.groupby('year_month').median()

plt.plot(data_mean.index,data_mean.approve_percent, 'red')

data_mean.index provides the year's month column, treated as an index column in the

data_mean data frame.

plt.plot(data_median.index, data_median.approve_percent, 'green')

plt.legend(['Mean', 'Median'])

These data_mean and median lines can be shown along side the original scatter graph data

points by plotting:

plt.plot(data.year_month, data.approve_percent, 'o', markersize=2, alpha=0.3)

plt.show()

Ploting the main graph:

To find the 25th and 75th percentiles:

data_25 = data.groupby('year_month').quantile(0.25) data_75 = data.groupby('year_month').quantile(0.75)

fig = plt.figure(figsize=(18, 12), dpi= 80, facecolor='w', edgecolor='k')

plt.plot(data_75.index, data_75.approve_percent, 'green') plt.plot(data_median.index, data_median.approve_percent, 'red') plt.plot(data_25.index, data_25.approve_percent, 'blue') plt.plot(data.year_month, data.approve_percent, 'o', markersize=2, alpha=0.8)

plt.legend(['75th Percentile', 'Median', '25th Percentile']) plt.xlabel('Y E A R', fontsize='20', color='forestgreen') plt.ylabel('A P P R O V A L %', fontsize='20', color='forestgreen') plt.title('O B A M A 'S A P P R O V A L R A T I N G %', fontsize='28', color='forestgreen') plt.show()