0

I essentially would like to replicate this image from Jiao2020:

enter image description here

Having a glimpse at my data we have:

    day         hour    consumption
0   Tuesday     17         -0.003196
1   Saturday    8           0.000000
2   Wednesday   17          0.000000
3   Tuesday     16         -0.002542
4   Tuesday     15          0.000000


RangeIndex: 402 entries, 0 to 401
Data columns (total 3 columns):
 #   Column       Non-Null Count  Dtype   
---  ------       --------------  -----   
 0   day_start    402 non-null    category
 1   hour_start   402 non-null    int64   
 2   consumption  402 non-null    float64

I'd like to have day on the y-axis, hour on the x-axis, and consumption as the value being measured.

To plot it I'm simply trying:

sns.heatmap(df)
plt.show()

but I get the error:

ValueError: could not convert string to float: 'Tuesday'

I've seen posts to leave the day as a da number instead of string, but I'd like to show the days' names on the plot instead of their numbers. How can this plot be achieved?

NOTE: My original timestamp (before extracting information from it) is given in the format "%Y-%m-%d %H:%M:%S.%f". For example, 2021-06-15 17:49:57.828.

Joehat
  • 709
  • 6
  • 18
  • 1
    You need to create a pivot, similar to the flights example at seaborn's [`sns.heatmap` page](https://seaborn.pydata.org/generated/seaborn.heatmap.html). Something like `sns.heatmap(df.pivot("day", "hour", "consumption"))` – JohanC Sep 09 '21 at 18:37

0 Answers0