Another tutorial done under the concept of “geospatial python”. The tutorial shows the procedure to run a Scipy interpolation over a Pandas dataframe of point related data having a 2D Numpy array as an output. With some procedures of Rasterio the Numpy array was transformed into a monoband geospatial Tiff raster.

Tutorial

Code

import numpy as np
import pandas as pd
from scipy.interpolate import griddata
import matplotlib.pyplot as plt
import rasterio

#open the chemistry data
chemData = pd.read_csv('../tab/chemistryData.csv',index_col=0,usecols=[0,5,6,18])
chemData = chemData.dropna() #delete missing data rows
chemData.describe()

	EsteCor	NorteCor	pH
count	176.000000	1.760000e+02	176.000000
mean	247362.431818	8.352024e+06	8.070892
std	5246.316476	7.225883e+03	0.877716
min	236536.000000	8.336049e+06	5.630000
25%	242377.750000	8.346304e+06	7.617500
50%	249067.000000	8.352004e+06	8.045000
75%	251132.000000	8.356741e+06	8.510000
max	259877.000000	8.372586e+06	10.400000

# Optional : drop extreme values
chemData = chemData[chemData.pH > np.quantile(chemData.pH,0.1)]
chemData = chemData[chemData.pH < np.quantile(chemData.pH,0.9)]
chemData.describe()

	EsteCor	NorteCor	pH
count	140.000000	1.400000e+02	140.000000
mean	247838.157143	8.352633e+06	8.076693
std	5230.177137	7.042238e+03	0.489108
min	236536.000000	8.336049e+06	7.100000
25%	243042.750000	8.347914e+06	7.790000
50%	250146.000000	8.353390e+06	8.055000
75%	251441.250000	8.357699e+06	8.370000
max	259877.000000	8.372586e+06	9.330000

#define interpolation inputs
points = list(zip(chemData.EsteCor,chemData.NorteCor))
values = chemData.pH.values
print(points[:5])
print(values[:5])

[(250901.0, 8342712.0), (244329.0, 8346119.0), (244329.0, 8346119.0), (244329.0, 8346119.0), (240476.0, 8351811.000000001)]
[8.03 8.96 7.8  8.04 8.17]

#define raster resolution
rRes = 50

#create coord ranges over the desired raster extension
xRange = np.arange(chemData.EsteCor.min(),chemData.EsteCor.max()+rRes,rRes)
yRange = np.arange(chemData.NorteCor.min(),chemData.NorteCor.max()+rRes,rRes)
print(xRange[:5],yRange[:5])

[236536. 236586. 236636. 236686. 236736.] [8336049. 8336099. 8336149. 8336199. 8336249.]

#create arrays of x,y over the raster extension
gridX,gridY = np.meshgrid(xRange, yRange)

#interpolate over the grid
gridPh = griddata(points, values, (gridX,gridY), method='linear')

#show interpolated values
plt.imshow(gridPh)

#definition of the raster transform array
from rasterio.transform import Affine
transform = Affine.translation(gridX[0][0]-rRes/2, gridY[0][0]-rRes/2)*Affine.scale(rRes,rRes)
transform

Affine(50.0, 0.0, 236511.0,
       0.0, 50.0, 8336023.999999999)

#get crs as wkt
from rasterio.crs import CRS
rasterCrs = CRS.from_epsg(32718)
rasterCrs.data

{'init': 'epsg:32718'}

#definition, register and close of interpolated raster
interpRaster = rasterio.open('../rst/interpRaster3.tif',
                                'w',
                                driver='GTiff',
                                height=gridPh.shape[0],
                                width=gridPh.shape[1],
                                count=1,
                                dtype=gridPh.dtype,
                                crs=rasterCrs,
                                transform=transform,
                                )
interpRaster.write(gridPh,1)
interpRaster.close()