CODEWITHSUNDEEP
pythonpandasmatplotlib
top bantz
top bantz

Reputation: 615

Color different values in a Pandas dataframe a different color on a scatterpot with Matplotlib

I have a df such as this:

Date                   Sig
2020-06-16 09:00:00    1
2020-06-16 13:00:00    0
2020-06-16 17:00:00   -1
2020-06-16 21:00:00   -1
2020-06-17 01:00:00    0
2020-06-17 05:00:00    1
2020-06-17 09:00:00    1
2020-06-17 13:00:00    0
2020-06-17 17:00:00    0
2020-06-17 21:00:00    -1

What I'm trying to do is a simple scatter plot in Matplotlib with the index as the X axis and the Sig column on the Y, but only showing 1s and -1s, with 1 colored blue and -1 colored red. But I don't want a 0 row to be omitted from the scatter plot if that makes sense.

I can obviously just do the scatter plot thusly 'plt.scatter(x=df.index, y=df['Sig'])' but that of course plots all numbers in Sig and doesn't color code them.

Any help would be really appreciated. Cheers

Upvotes: 0

Views: 425

Answers (2)

Guimoute
Guimoute

Reputation: 4629

plt.scatter supports iterables of colors, so you can actually give each point an independent color.

Here, we will simply have a dictionary with three choices dictated by the sign of your data:

colors = {-1: "red", 0: "black", +1: "blue"}
plt.scatter(x=df.index, y=df["Sig"], color=[colors[sign] for sign in df["Sig"]])

Upvotes: 1

Quang Hoang
Quang Hoang

Reputation: 150745

You can do a groupby on filtered data like this:

# datetime data type is recommended
df['Date'] = pd.to_datetime(df['Date'])

colors = ['r','b']
fig, ax = plt.subplots()
for c, (k,d) in zip(colors, df[df['Sig']!=0].groupby('Sig')):
    d.plot.scatter(x='Date',y='Sig',label=k, ax=ax, c=c)

Output:

enter image description here

Upvotes: 1

Related Questions