Is the DataFrame['2002':'2005'][['Value1','Value2']].bar(any args?) possible way to create a bar plot of Dataframe, where 2 values are distributed within a long period of time.
I can create a simple plot, but I want bars (1 bar - 1 day).
If there is no such a simple way, what would be the simpliest one?
Well, not too self coding, but anyway, possible, so it works... (thanks to these guys here...)
Suppose you have a dataframe like
df.head()
temperature pressure
2018-10-01 21.860016 1031.418143
2018-10-02 20.590761 1063.008550
2018-10-03 21.356381 1047.183300
2018-10-04 20.393329 1037.710172
2018-10-05 20.716377 1027.680324
... ... ...
Then you'd need to import
import matplotlib.pyplot as plt
import matplotlib.dates as md
to plot your data like
fig, ax = plt.subplots()
ax.xaxis_date()
ax.bar(df.index -md.num2timedelta(.2), df.temperature, .4, color='r', label = df.temperature.name)
ax2 = ax.twinx()
ax2.bar(df.index +md.num2timedelta(.2), df.pressure, .4, color='b', label = df.pressure.name)
ax.xaxis.set_major_formatter(md.DateFormatter('%b %d'))
ax.xaxis.set_major_locator(md.WeekdayLocator())
fig.legend()
Result:
Related
This question already has answers here:
Annotate bars with values on Pandas bar plots
(4 answers)
Closed 1 year ago.
I would like to create an annotation to a bar chart that compares the value of the bar to two reference values. An overlay such as shown in the picture, a kind of staff gauge, is possible, but I'm open to more elegant solutions.
The bar chart is generated with the pandas API to matplotlib (e.g. data.plot(kind="bar")), so a plus would be if the solution is playing nicely with that.
You may use smaller bars for the target and benchmark indicators. Pandas cannot annotate bars automatically, but you can simply loop over the values and use matplotlib's pyplot.annotate instead.
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
a = np.random.randint(5,15, size=5)
t = (a+np.random.normal(size=len(a))*2).round(2)
b = (a+np.random.normal(size=len(a))*2).round(2)
df = pd.DataFrame({"a":a, "t":t, "b":b})
fig, ax = plt.subplots()
df["a"].plot(kind='bar', ax=ax, legend=True)
df["b"].plot(kind='bar', position=0., width=0.1, color="lightblue",legend=True, ax=ax)
df["t"].plot(kind='bar', position=1., width=0.1, color="purple", legend=True, ax=ax)
for i, rows in df.iterrows():
plt.annotate(rows["a"], xy=(i, rows["a"]), rotation=0, color="C0")
plt.annotate(rows["b"], xy=(i+0.1, rows["b"]), color="lightblue", rotation=+20, ha="left")
plt.annotate(rows["t"], xy=(i-0.1, rows["t"]), color="purple", rotation=-20, ha="right")
ax.set_xlim(-1,len(df))
plt.show()
There's no direct way to annotate a bar plot (as far as I am aware) Some time ago I needed to annotate one so I wrote this, perhaps you can adapt it to your needs.
import matplotlib.pyplot as plt
import numpy as np
ax = plt.subplot(111)
ax.set_xlim(-0.2, 3.2)
ax.grid(b=True, which='major', color='k', linestyle=':', lw=.5, zorder=1)
# x,y data
x = np.arange(4)
y = np.array([5, 12, 3, 7])
# Define upper y limit leaving space for the text above the bars.
up = max(y) * .03
ax.set_ylim(0, max(y) + 3 * up)
ax.bar(x, y, align='center', width=0.2, color='g', zorder=4)
# Add text to bars
for xi, yi, l in zip(*[x, y, list(map(str, y))]):
ax.text(xi - len(l) * .02, yi + up, l,
bbox=dict(facecolor='w', edgecolor='w', alpha=.5))
ax.set_xticks(x)
ax.set_xticklabels(['text1', 'text2', 'text3', 'text4'])
ax.tick_params(axis='x', which='major', labelsize=12)
plt.show()
In Pandas, I am trying to generate a Ridgeline plot for which the density values are shown (either as Y axis or color-ramp). I am using the Joyplot but any other alternative ways are fine.
So, first I created the Ridge plot to show the different distribution plot for each condition (you can reproduce it using this code):
import pandas as pd
import joypy
import matplotlib
import matplotlib.pyplot as plt
df1 = pd.DataFrame({'Category1':np.random.choice(['C1','C2','C3'],1000),'Category2':np.random.choice(['B1','B2','B3','B4','B5'],1000),
'year':np.arange(start=1900, stop=2900, step=1),
'Data':np.random.uniform(0,1,1000),"Period":np.random.choice(['AA','CC','BB','DD'],1000)})
data_pivot=df1.pivot_table('Data', ['Category1', 'Category2','year'], 'Period')
fig, axes = joypy.joyplot(data_pivot, column=['AA', 'BB', 'CC', 'DD'], by="Category1", ylim='own', figsize=(14,10), legend=True, alpha=0.4)
so it generates the figure but without my desired Y axis. So, based on this post, I could add a colorramp, which neither makes sense nor show the differences between the distribution plot of the different categories on each line :) ...
ar=df1['Data'].plot.kde().get_lines()[0].get_ydata() ## a workaround to get the probability values to set the colorramp max and min
norm = plt.Normalize(ar.min(), ar.max())
original_cmap = plt.cm.viridis
cmap = matplotlib.colors.ListedColormap(original_cmap(norm(ar)))
sm = matplotlib.cm.ScalarMappable(cmap=original_cmap, norm=norm)
sm.set_array([])
# plotting ....
fig, axes = joypy.joyplot(data_pivot,colormap = cmap , column=['AA', 'BB', 'CC', 'DD'], by="Category1", ylim='own', figsize=(14,10), legend=True, alpha=0.4)
fig.colorbar(sm, ax=axes, label="density")
But what I want is some thing like either of these figures (preferably with colorramp) :
I have a DataFrame and need to make a scatter-plot from it.
I need to use 2 columns as the x-axis and y-axis and only need to plot 2 rows from the entire dataset. Any suggestions?
For example, my dataframe is below (50 states x 4 columns). I need to plot 'rgdp_change' on the x-axis vs 'diff_unemp' on the y-axis, and only need to plot for the states, "Michigan" and "Wisconsin".
So from the dataframe, you'll need to select the rows from a list of the states you want: ['Michigan', 'Wisconsin']
I also figured you would probably want a legend or some way to differentiate one point from the other. To do this, we create a colormap assigning a different color to each state. This way the code is generalizable for more than those two states.
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.cm as cm
import matplotlib.colors as colors
# generate a random df with the relevant rows, columns to your actual df
df = pd.DataFrame({'State':['Alabama', 'Alaska', 'Michigan', 'Wisconsin'], 'real_gdp':[1.75*10**5, 4.81*10**4, 2.59*10**5, 1.04*10**5],
'rgdp_change': [-0.4, 0.5, 0.4, -0.5], 'diff_unemp': [-1.3, 0.4, 0.5, -11]})
fig, ax = plt.subplots()
states = ['Michigan', 'Wisconsin']
colormap = cm.viridis
colorlist = [colors.rgb2hex(colormap(i)) for i in np.linspace(0, 0.9, len(states))]
for i,c in enumerate(colorlist):
x = df.loc[df["State"].isin(['Michigan', 'Wisconsin'])].rgdp_change.values[i]
y = df.loc[df["State"].isin(['Michigan', 'Wisconsin'])].diff_unemp.values[i]
legend_label = states[i]
ax.scatter(x, y, label=legend_label, s=50, linewidth=0.1, c=c)
ax.legend()
plt.show()
Use the dataframe plot method, but first filter the sates you need using index isin method:
states = ["Michigan", "Wisconsin"]
df[df.index.isin(states)].plot(kind='scatter', x='rgdp_change', y='diff_unemp')
TL;DR: In pandas how do I plot a bar chart so that its x axis tick labels look like those of a line chart?
I made a time series with evenly spaced intervals (one item each day) and can plot it like such just fine:
intensity[350:450].plot()
plt.show()
But switching to a bar chart created this mess:
intensity[350:450].plot(kind = 'bar')
plt.show()
I then created a bar chart using matplotlib directly but it lacks the nice date time series tick label formatter of pandas:
def bar_chart(series):
fig, ax = plt.subplots(1)
ax.bar(series.index, series)
fig.autofmt_xdate()
plt.show()
bar_chart(intensity[350:450])
Here's an excerpt from the intensity Series:
intensity[390:400]
2017-03-07 3
2017-03-08 0
2017-03-09 3
2017-03-10 0
2017-03-11 0
2017-03-12 0
2017-03-13 2
2017-03-14 0
2017-03-15 3
2017-03-16 0
Freq: D, dtype: int64
I could go all out on this and just create the tick labels by hand completely but I'd rather not have to baby matplotlib and let do pandas its job and do what it did in the very first figure but with a bar plot. So how do I do that?
Pandas bar plots are categorical plots. They create one tick (+label) for each category. If the categories are dates and those dates are continuous one may aim at leaving certain dates out, e.g. to plot only every fifth category,
ax = series.plot(kind="bar")
ax.set_xticklabels([t if not i%5 else "" for i,t in enumerate(ax.get_xticklabels())])
In contrast, matplotlib bar charts are numberical plots. Here a useful ticker can be applied, which ticks the dates weekly, monthly or whatever is needed.
In addition, matplotlib allows to have full control over the tick positions and their labels.
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from matplotlib import dates
index = pd.date_range("2018-01-26", "2018-05-05")
series = pd.Series(np.random.rayleigh(size=100), index=index)
plt.bar(series.index, series.values)
plt.gca().xaxis.set_major_locator(dates.MonthLocator())
plt.gca().xaxis.set_major_formatter(dates.DateFormatter("%b\n%Y"))
plt.show()
I'm trying to plot timeseries revenue data by quarter with matplotlib.pyplot but keep getting an error. Below is my code and the errors The desired behavior is to plot the revenue data by quarter using matplotlib. When I try to do this, I get:
TypeError: Axis must havefreqset to convert to Periods
Is it because timeseries dates expressed as periods cannot be plotted in matplotlib? Below is my code.
def parser(x):
return pd.to_datetime(x, format='%m%Y')
tot = pd.read_table('C:/Desktop/data.txt', parse_dates=[2], index_col=[2], date_parser=parser)
tot = tot.dropna()
tot = tot.to_period('Q').reset_index().groupby(['origin', 'date'], as_index=False).agg(sum)
tot.head()
origin date rev
0 KY 2016Q2 1783.16
1 TN 2014Q1 32128.36
2 TN 2014Q2 16801.40
3 TN 2014Q3 33863.39
4 KY 2014Q4 103973.66
plt.plot(tot.date, tot.rev)
If you want to use matplotlib, the following code should give you the desired plot:
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
df = pd.DataFrame({'origin': ['KY','TN','TN','TN','KY'],
'date': ['2016Q2','2014Q1','2014Q2','2014Q3','2014Q4'],
'rev': [1783.16, 32128.36, 16801.40, 33863.39, 103973.66]})
x = np.arange(0,len(df),1)
fig, ax = plt.subplots(1,1)
ax.plot(x,df['rev'])
ax.set_xticks(x)
ax.set_xticklabels(df['date'])
plt.show()
You could use the xticks command and represent the data with a bar chart with the following code:
plt.bar(range(len(df.rev)), df.rev, align='center')
plt.xticks(range(len(df.rev)), df.date, size='small')
It seems like bug.
For me works DataFrame.plot:
ooc.plot(x='date', y='rev')