}
# names of statistical variables
metrics = "sum mean std".split()
# Specify columns to join from shapefile
shp_cols =shp.drop(['geometry'],axis=1)
# Read rasters and join names with path
rasters = {}
for n, d in dirs.items():
raster = (os.path.join(d, n+".tif"))
rasters[n]= raster
#copy metrics
spfeas_stats = shp_cols.copy()
# Calculate Rasters
for rast, path in rasters.items():
print rast
stats = zs(shp, path, nodata=-999, stats=metrics)
new_colnames = ["{}_{}".format(rast, metric) for metric in metrics]
df = pd.DataFrame(stats)
df2 = df.rename(columns=dict(zip(metrics, new_colnames)))
spfeas_stats =spfeas_stats.join(df2)
# Save dataframe as csv
spfeas_stats.to_csv("hog_stats_all.csv")
import rasterio as rio
from rasterstats import zonal_stats as zs
import geopandas as gpd
import pandas as pd
import os
from glob import glob
#Read shapefile
shp = gpd.GeoDataFrame.from_file("../zonal_stats/GN_Divisions.shp")
data_dir ="Sri_Lanka_lac_bands"
rasters =glob(os.path.join(data_dir, "*.tif"))
for i in rasters:
with rio.open(i) as src:
transform=src.meta['transform']
array = src.read()
metrics = "sum mean std"
shp_cols =shp[['gid', 'gnd_c','gnd_n']]
#Run zonal stat and save values as csv
for i in array:
stats = zs(shp, i, stats = metrics, transform=transform, prefix ="lac_")
pd_stats =shp_cols.join(pd.DataFrame(stats))
pd_stats.to_csv('lac_all_stats.csv')
# Add file name and relative path to dictionary
rasters[file_name] = tif_relative_path
# Set output location and file name for zonal statistics CSV file *
zonal_stats_location = os.path.join(output_directory, "zonal_stats",
"zonalstats_{}.csv".format(feature))
# Copy metrics
spfeas_stats = shp_cols.copy()
# Calculate Rasters
for rast, path in rasters.items():
# Calculate stats, this creates a dictionary with metrics as columns and values as rows
stats = zs(shp, path, stats=metrics)
# New column names using formatter
new_colnames = ["{}_{}".format(rast, metric) for metric in metrics]
# Create dataframe from stats dictionary
df = pd.DataFrame(stats)
# Rename the metrics column with the new column names
# Zip puts the indices of two lists together (zip them together)
# Dict creates a dictionary because rename needs a dictionary (old name: new name)
df2 = df.rename(columns=dict(zip(metrics, new_colnames)))
spfeas_stats = spfeas_stats.join(df2)
# Save dataframe as csv
spfeas_stats.to_csv(zonal_stats_location)
raster7="orb_sc51_mean.tif"
raster8="orb_sc51_skew.tif"
raster9="orb_sc51_kurtosis.tif"
raster10="orb_sc51_variance.tif"
raster11="orb_sc71_max.tif"
raster12="orb_sc71_mean.tif"
raster13="orb_sc71_skew.tif"
raster14="orb_sc71_kurtosis.tif"
raster15="orb_sc71_variance.tif"
stats1 = zs(shp, raster1, stats =
#specify zonal stat mettrics
metrics = "sum mean std"
#Run zonal stat and save values as csv
stats1 = zs(shp, raster1, stats = metrics, prefix ="lac_sc5_")
stats2 = zs(shp, raster2, stats = metrics, prefix ="lac_sc7_")
stats3 =
pd_stats = pd.DataFrame(stats1).join(pd.DataFrame(stats2))
stats_final = shp.join(pd_stats)
stats_final.to_csv('lac_stats.csv')
max_depth.read_masks(1)),
transform=max_depth.transform)
],
crs=max_depth.crs)
# Store original areas for damage calculation
buildings['original_area'] = buildings.area
# Buffer buildings
buildings['geometry'] = buildings.buffer(buffer)
# Extract maximum depth and vd_product for each building
buildings['depth'] = [
row['max'] for row in zs(buildings,
depth,
affine=max_depth.transform,
stats=['max'],
all_touched=True,
nodata=max_depth.nodata)
]
# Filter buildings
buildings = buildings[buildings['depth'] > threshold]
# Calculate depth above floor level
buildings['depth'] = buildings.depth - threshold
if len(buildings) == 0:
with open(os.path.join(outputs_path, 'buildings.csv'), 'w') as f:
f.write('')
exit(0)
'mean_sc7_mean': 'Sri_Lanka_mean_bands/',
'mean_sc7_variance': 'Sri_Lanka_mean_bands/'
}
# names of statistical variables
metrics = "sum mean std".split()
# Read rasters and join names with path
rasters = {}
for n, d in dirs.items():
for r in metrics:
raster = (os.path.join(d, n + ".tif"))
name = n + "_" + r
rasters[name] = raster
# Specify columns to join from shapefile
shp_cols = shp.drop(['geometry'], axis=1)
# Calculate Rasters
for raster, path in rasters.items():
print raster
stats = zs(shp, path, nodata=0, stats=metrics)
df = pd.DataFrame(stats)
# Rename names of columns
new_colnames = ["{}_{}".format(raster, metric) for metric in metrics]
df2 = df.rename(columns=dict(zip(metrics, new_colnames)))
working_zones = shp_cols.join(df2) # append to working copy
# Save dataframe as csv
working_zones.to_csv("means_stats_all.csv")
import rasterio as rio
from rasterstats import zonal_stats as zs
import geopandas as gpd
import pandas as pd
import os
#Read shapefile
shp = gpd.GeoDataFrame.from_file("../../zonal_stats/GN_Divisions.shp")
#Read Raster
raster1="lac_sc5_lac.tif"
raster2="lac_sc7_lac.tif"
#specify zonal stat mettrics
metrics = "sum mean std"
#Run zonal stat and save values as csv
stats1 = zs(shp, raster1, stats = metrics, prefix ="lac_sc5_")
stats2 = zs(shp, raster2, stats = metrics, prefix ="lac_sc7_")
pd_stats = pd.DataFrame(stats1).join(pd.DataFrame(stats2))
stats_final = shp.join(pd_stats)
stats_final.to_csv('lac_stats.csv')