Skip to content
Advertisement

How to plot a dataframe column of lists as horizontal lines

I have a Dataframe with the column 'all_maxs' that could have a list of different values.

          c            all_maxs
38  50804.6           [50883.3]
39  50743.9           [50883.3]
40  50649.9           [50883.3]
41  50508.3           [50883.3]
42  50577.6           [50883.3]
43  50703.0           [50883.3]
44  50793.7           [50883.3]
45  50647.8  [50883.3, 50813.1]
46  50732.8  [50883.3, 50813.1]
47  50673.2  [50883.3, 50813.1]

df.plot(y='c')

Current Result

enter image description here

I need to plot column 'c', and the values of column 'all_maxs' that should be horizontal lines.

Expected Result

expected result

Advertisement

Answer

  1. Verify the 'all_maxs' values are list type.
  2. Extract values from the lists and plot them as horizontal lines.
    • df = df.dropna() if there are any NaN

Imports and DataFrame

  • Convert the 'all_maxs' column type from str to list if needed, using ast.liter_eval
import pandas as pd
from ast import literal_eval

data =
{38: {'all_maxs': '[50883.3]', 'c': 50804.6},
 39: {'all_maxs': '[50883.3]', 'c': 50743.9},
 40: {'all_maxs': '[50883.3]', 'c': 50649.9},
 41: {'all_maxs': '[50883.3]', 'c': 50508.3},
 42: {'all_maxs': '[50883.3]', 'c': 50577.6},
 43: {'all_maxs': '[50883.3]', 'c': 50703.0},
 44: {'all_maxs': '[50883.3]', 'c': 50793.7},
 45: {'all_maxs': '[50883.3, 50813.1]', 'c': 50647.8},
 46: {'all_maxs': '[50883.3, 50813.1]', 'c': 50732.8},
 47: {'all_maxs': '[50883.3, 50813.1]', 'c': 50673.2}}

df = pd.DataFrame.from_dict(data, orient='index')

# reorder the columns to match the OP
df = df[['c', 'all_maxs']]

# print a value from all_maxs to see the type
>>> print(type(df.loc[38, 'all_maxs']))
str

# currently the all_max values are strings, which must be converted to list type
df.all_maxs = df.all_maxs.apply(literal_eval)

# print a value from all_maxs to see the type
>>> print(type(df.loc[38, 'all_maxs']))
list

Plot

  • Plot the dataframe directly with pandas.DataFrame.plot
    • xticks=df.index will make an xtick for every value in the index, but if there are many values crowding the x-axis, remove this parameter.
  • Use matplotlib.pyplot.hlines, which will accept a list of values, to plot the unique values from 'all_max' as horizontal lines.
    • Use pandas.DataFrame.explode to remove all the values from lists, and then drop duplicates with .drop_duplicates
    • y= will be the remaining values in the 'all_maxs' column
    • xmin= will be the remaining index values
    • xmax= will be the max value in the index plotted from df
ax = df.plot(y='c', legend=False, figsize=(8, 5), xticks=df.index)

# extract all the values from all_maxs, drop the duplicates
all_maxs = df.all_maxs.explode().drop_duplicates().to_frame()

# add the horizontal lines
ax.hlines(y=all_maxs.all_maxs, xmin=all_maxs.index, xmax=df.index.max(), color='k')

enter image description here

Advertisement