Seaborn Heatmap with single column

Question

I have a dataframe that has an index (words) and a single column (counts) for some lyrics. I am trying to create a heatmap based on the word counts.

    Cuenta
Que 179
La  145
Y   142
Me  113
No  108

I am trying to produce the heatmap like this:

df1 = pd.DataFrame.from_dict([top50]).T
df1.columns = ['Cuenta']
df1.sort_values(['Cuenta'], ascending = False, inplace=True)

result = df1.pivot(index=df1.index, columns='Cuenta', values=df1.Cuenta.count)
sns.heatmap(result, annot=True, fmt="g", cmap='viridis')
plt.show()

But, it keeps throwing 'Index' object has no attribute 'levels'

Any ideas why this isn't working? I tried using the index or words as a separate column and still doesn't work.

Answer 1

The data is one-dimensional. The counts are already present in the one (and only) column of the dataframe. There is no meaningless way to pivot this data.

You would hence directly plot the dataframe as a heatmap.

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

df = pd.DataFrame({"Cuenta": [179,145,142,113,108]},
                  index=["Que", "La", "Y", "Me", "No"])

sns.heatmap(df, annot=True, fmt="g", cmap='viridis')

plt.show()

---

• If the data to be on the y-axis is a column, and not the index of the dataframe, then use .set_index

df = pd.DataFrame({"Cuenta": [179,145,142,113,108],
                   "words": ["Que", "La", "Y", "Me", "No"]})

# given a dataframe of two columns, set the column as the index
df.set_index("words", inplace=True)

ax = sns.heatmap(df, annot=True, fmt="g", cmap='viridis')

---

• sns.heatmap will result in an IndexError if passing a pandas.Series.
  • .value_counts create a Series
  • df['column'] and df.column create a Series. Use df[['column']] instead.

# sample data
tips = sns.load_dataset('tips')

# value_counts creates a Series
vc = tips.time.value_counts()

# convert to a DataFrame
vc = vc.to_frame()

# plot
ax = sns.heatmap(data=vc)