How to specify different color for a specific year value range in a single figure? (Python)

Question

I've a time-series dataset, from 1992-2017. I can set a color for the whole data dots but what I want is to set desired color for specific year range. For Example; from 1992-1995 "Blue", from 1995-2005 "Red" etc. How can we do that?

Dataset has 2 columns; year and value.

import numpy as np
import pandas as pd
from scipy import stats
from sklearn import linear_model
from matplotlib import pyplot as plt
import pylab
import matplotlib.patches as mpatches
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.linear_model import LinearRegression

Atlantic = pd.read_csv('C:\\AtlanticEnd.csv', error_bad_lines=False)

X = Atlantic['year']

y = Atlantic['Poseidon']

plt.figure(figsize=(20,10))
plt.ylabel('Change in mean sea level [mm]', fontsize=20)
plt.xlabel('Years', fontsize=20)
plt.title('Atlantic Ocean - Mean Sea Level', fontsize=20)
colors = ["blue", "red", "green", "purple"]
texts = ["Poseidon", "Jason1", "Jason2", "Jason3"]
patches = [ plt.plot([],[], marker="o", ms=10, ls="", mec=None, color=colors[i], 
            label="{:s}".format(texts[i]) )[0]  for i in range(len(texts)) ]
plt.legend(handles=patches, loc='upper left', ncol=1, facecolor="grey", numpoints=1 )

plt.plot(X, y, 'ro', color='red')

slope, intercept, r_value, p_value, std_err = stats.linregress(X, y)
plt.plot(X, X*slope+intercept, 'b')

plt.axis([1992, 2018, -25, 80])

plt.grid(True)

plt.show()

def trendline(Atlantic, order=1):
    coeffs = np.polyfit(Atlantic.index.values, list(Atlantic), order)
    slope = coeffs[-2]
    return float(slope)

slope = trendline(y)
print(slope)

enter image description here

Answer 1

I could imagine that using a colormap for a scatter plot of the points may be an easy solution. The scatter's color would then simply be defined by the year, assuming the year is given in decimal format. A BoundaryNorm would define the ranges for the values and a colormap can easily be created from a list of colors.

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.colors

y =  np.random.rand(300)*26+1992
d = (3.075*(y-1992)-17)+np.random.normal(0,5,300)
df = pd.DataFrame({"year" : y, "data" : d})

bounds = [1992,1995,2005,2015,2018]
colors = ["darkorchid", "crimson", "limegreen", "gold"]
cmap = matplotlib.colors.ListedColormap(colors)
norm = matplotlib.colors.BoundaryNorm(bounds, len(colors))

fig, ax = plt.subplots()
sc = ax.scatter(df.year, df.data, c=df.year.values, cmap=cmap, norm=norm)
fig.colorbar(sc, spacing="proportional")

fit = np.polyfit(df.year.values, df.data.values, deg=1)
ax.plot(df.year, np.poly1d(fit)(df.year.values), color="k")

plt.show()

Answer 2

I made my own random data for this function to work but assuming you have non-overlapping date ranges, this should work. It also seemed like your dates are not of pd.datetime type. This should work for pd.datetime types but your lookup values in the dictionary will be something like ("1992-01-01","2000-01-01") and so on.

# Create data
data = np.random.rand(260,1)
dates = np.array(list(range(1992,2018))*10)

df = pd.DataFrame({"y":data[:,0],"date":dates})
df = df.sort(columns="date")

# Dictionary lookup
lookup_dict = {(1992,2000):"r", (2001,2006):"b",(2007,2018):"k"}

# Slice data and plot
fig, ax = plt.subplots()
for lrange in lookup_dict:
    temp = df[(df.date>=lrange[0]) & (df.date<=lrange[1])]
    ax.plot(temp.date,temp.y,color=lookup_dict[lrange], marker="o",ls="none")

This produces: