0

I'm relatively new to python and I'm having some issues trying to plot some data with pandas - please help! The relevant sections of my code are below.

I create a new data-frame, do some analysis with it, then split it into two separate data-frames (one for 2016 data, one for 2017 data). I then calculate some new columns for the 2017 data-frame (percentage difference between these values and the values for 2016). Finally, I make several plots to show this data.

I'm trying to modify the x axis tick labels to show each month as e.g. '07/17' for July 2017. I'm using a method I've used before, however it really doesn't seem to be working! 

import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
import matplotlib as mp
import numpy as np
import seaborn as sns

# set up index dates
dates1 = pd.date_range(start='2016-06-01', end='2016-12-01', freq='MS')
dates2 = pd.date_range(start='2017-06-01', end='2017-12-01', freq='MS')
dates = dates1.union(dates2)

# create data-frame
data = {'sessions':[609981,734725,713928,602909,562960,425774,394887,1094831,
                    1210158,1141875,898867,798469,642942,546186],
        'conversion':[5.40,4.98,4.89,4.68,3.99,3.54,3.71,4.17,4.23,4.13,4.3,3.86,3.24,3.60],
       'avg_value':[98.12,95.43,89.94,86.99,83.23,86.31,93.34,98.01,96.89,90.82,86.00,84.40,85.90,89.53]}
df = pd.DataFrame(data,index=dates)

# calculate revenue and yearly changes
df['revenue'] = (df.sessions*df.conversion/100.)*df.avg_value

df_2016 = df['Jan 2016':'Dec 2016'].copy()
df_2017 = df['Jan 2017':'Dec 2017'].copy()

df_2017['d_avg_value'] = (df_2017.avg_value.values - df_2016.avg_value.values)*100 / df_2016.avg_value.values
df_2017['d_conversion'] = (df_2017.conversion.values - df_2016.conversion.values)*100 / df_2016.conversion.values
df_2017['d_sessions'] = (df_2017.sessions.values - df_2016.sessions.values)*100 / df_2016.sessions.values
df_2017['d_revenue'] = (df_2017.revenue.values - df_2016.revenue.values)*100 / df_2016.revenue.values

# set up subplot layout
fig, axes = plt.subplots(figsize=(10, 10), nrows=2, ncols=2)
plt.subplots_adjust(wspace=0.3, hspace=0.3)

# populate subplots
df_2017.plot(ax=axes[0,0], y='d_sessions',marker='o',linestyle='none', legend=None)
df_2017.plot(ax=axes[0,1], y='d_conversion',marker='o',linestyle='none', legend=None)
df_2017.plot(ax=axes[1,0], y='d_avg_value',marker='o',linestyle='none', legend=None)
df_2017.plot(ax=axes[1,1], y='d_revenue',marker='o',linestyle='none', legend=None, color='red')

# change x-limits, x-tick labels
for ax in axes.reshape(-1):
    ax.set_xlim(['May 2017','Jan 2018'])
    ax.xaxis.set_major_locator(mdates.MonthLocator())
    ax.xaxis.set_major_formatter(mdates.DateFormatter('%m/%y'))

# label y-axes
axes[0,0].set_ylabel('% Change in Sessions')
axes[0,1].set_ylabel('% Change in Conversion')
axes[1,0].set_ylabel('% Change in Average Value')
axes[1,1].set_ylabel('% Change in Revenue')

plt.show()

enter image description here

astro person
  • 381
  • 5
  • 16
  • I've looked at that 'duplicate' question, but in the answers someone says that matplotlib and pandas datetimes are incompatible. If this is the issue then I don't understand how I was able to successfully use this method before (on a different set of plots, dataframe df)? – astro person Mar 11 '18 at 19:02
  • It depends, but if your question is why it works in one case and not the other you need to present both cases inside the question. Otherwise the general answer is indeed that they are incompatible. – ImportanceOfBeingErnest Mar 11 '18 at 19:23
  • Answer: I needed to add the "x_compat=True" keyword to my calls to the pandas plotting method. I think this addresses the incompatibility issue. – astro person Mar 11 '18 at 19:25
  • Oh well, indeed. I updated the solution. – ImportanceOfBeingErnest Mar 11 '18 at 19:38

0 Answers0