Back to Article
Data Screening
Download Notebook

Data Screening

Screening and filtering data from the IGN catalog down to the relevant events for La Palma
Author

Steve Purves

Load and review the data, check that dates are loaded properly and filter data down to the events in La Palma only. The dataset itself contains all recent IGN earthquake data for Spanish territory.

In [2]:
import pandas as pd
import matplotlib.pyplot as plt
In [3]:
df = pd.read_csv('../data/catalogoComunSV_1663233588717.csv', sep=';')
df.columns = ['Event', 'Date', 'Time', 'Latitude', 'Longitude', 'Depth(km)', 'Intensity','Magnitude', 'Type Mag', 'Location']
df['Time'] = df['Time'].str.strip()
df['Date'] = df['Date'].str.strip()
df['DateTime'] = (df['Date'] + ' ' + df['Time']).apply(pd.to_datetime, format='%d/%m/%Y %H:%M:%S')
df['Timestamp'] = pd.to_numeric(pd.to_datetime(df['DateTime']))
df['Date'] = pd.to_datetime(df['Date'], format="%d/%m/%Y")
df = df.sort_values(by=['DateTime'], ascending=True)
df.head()
Event Date Time Latitude Longitude Depth(km) Intensity Magnitude Type Mag Location DateTime Timestamp
0 es2017aaaga 2017-01-01 00:48:55 36.6765 -11.2736 0.0 2.8 6 SW CABO DE SAN VICENTE 2017-01-01 00:48:55 1483231735000000000
1 es2017aaakn 2017-01-01 01:28:17 28.1119 -16.2225 21.0 0.9 4 ATLÁNTICO-CANARIAS 2017-01-01 01:28:17 1483234097000000000
2 es2017aaang 2017-01-01 01:49:08 42.0648 -7.8471 22.0 2.0 4 SW RAIRIZ DE VEIGA.OU 2017-01-01 01:49:08 1483235348000000000
3 es2017aabdd 2017-01-01 02:36:20 38.3826 -9.3767 11.0 2.2 4 ATLÁNTICO-PORTUGAL 2017-01-01 02:36:20 1483238180000000000
4 es2017aabkh 2017-01-01 03:35:03 36.2492 -7.8227 13.0 1.9 4 GOLFO DE CÁDIZ 2017-01-01 03:35:03 1483241703000000000
In [4]:
df.describe()
Latitude Longitude Depth(km) Magnitude Type Mag Timestamp
count 58036.000000 58036.000000 58036.000000 58036.000000 58036.000000 5.803600e+04
mean 34.806854 -8.356855 11.450831 1.881846 4.031153 1.596182e+18
std 5.263327 7.119341 11.111965 0.763577 0.291332 4.954865e+16
min 26.349300 -19.930400 0.000000 -0.500000 3.000000 1.483232e+18
25% 28.563800 -16.672325 3.000000 1.400000 4.000000 1.559327e+18
50% 35.989200 -4.785250 10.000000 1.800000 4.000000 1.612319e+18
75% 37.890700 -3.234500 14.300000 2.400000 4.000000 1.635568e+18
max 44.985000 5.996600 115.000000 6.000000 6.000000 1.663232e+18
In [5]:
df.plot.scatter(x="Longitude", y="Latitude", figsize=(12,12), grid="on");

Filter down to La Palma events only

Also some rows seem to be missing depth measurements, (see differences in the count for columns above) - take only rows with valid depths

In [6]:
df = df[(df['Latitude'] < 28.7) & (df['Latitude'] > 28.3) & (df['Longitude'] < -17.7) & (df['Longitude'] > -18)]
df = df[df['Depth(km)'].notna()]
df.describe()
Latitude Longitude Depth(km) Magnitude Type Mag Timestamp
count 11347.000000 11347.000000 11347.000000 11347.000000 11347.0 1.134700e+04
mean 28.568893 -17.844062 14.089345 2.432423 4.0 1.633359e+18
std 0.021899 0.027345 8.624787 0.705282 0.0 2.007970e+16
min 28.310200 -17.995800 0.000000 0.200000 4.0 1.489103e+18
25% 28.558000 -17.854100 10.100000 1.900000 4.0 1.633777e+18
50% 28.564000 -17.838700 11.300000 2.600000 4.0 1.635671e+18
75% 28.574300 -17.829700 13.900000 2.900000 4.0 1.638294e+18
max 28.697800 -17.705700 46.900000 5.100000 4.0 1.663128e+18
In [7]:
plt.figure(figsize=(20,6))
plt.subplot(1,3,1)
df["Latitude"].plot.hist(bins=20, density=True)
df["Latitude"].plot.kde()
plt.subplot(1,3,2)
df["Longitude"].plot.hist(bins=20, density=True)
df["Longitude"].plot.kde()
plt.subplot(1,3,3)
df["Depth(km)"].plot.hist(bins=50, density=True)
_=df["Depth(km)"].plot.kde()

Spatial Plot

Scatter plot the spatial locations of events

In [14]:
from matplotlib import colormaps
cmap = colormaps['viridis_r']
ax = df.plot.scatter(x="Longitude", y="Latitude", 
                     s=40-df["Depth(km)"], c=df["Magnitude"], 
                     figsize=(12,10), grid="on", cmap=cmap)
colorbar = ax.collections[0].colorbar
colorbar.set_label("Magnitude")

plt.show()

A scatterplot of earthquake locations plotting latitude  against longitude.

Figure 1: Locations of earthquakes on La Palma since 2017.

Timeline Plot

Scatter plot the event time series and look for any quantization issues. Have times & dates been loaded correctly?

In [10]:
ax = df.plot.scatter(x='DateTime', y='Depth(km)', figsize=(20,8))
ax.set_ylim(50,0);

In [11]:
df['Swarm'] = 0
df.loc[(df['Date'] >= '2017-10-01') & (df['Date'] <= '2017-10-31'), 'Swarm'] = 1
df.loc[(df['Date'] >= '2019-02-01') & (df['Date'] <= '2019-02-28'), 'Swarm'] = 2
df.loc[(df['Date'] >= '2020-07-21') & (df['Date'] <= '2020-08-05'), 'Swarm'] = 3
df.loc[(df['Date'] >= '2020-10-04') & (df['Date'] <= '2020-10-10'), 'Swarm'] = 4
df.loc[(df['Date'] >= '2020-10-11') & (df['Date'] <= '2020-10-22'), 'Swarm'] = 5
df.loc[(df['Date'] >= '2020-11-15') & (df['Date'] <= '2020-11-29'), 'Swarm'] = 6
df.loc[(df['Date'] >= '2020-12-10') & (df['Date'] <= '2020-12-29'), 'Swarm'] = 7
df.loc[(df['Date'] >= '2020-12-10') & (df['Date'] <= '2020-12-31'), 'Swarm'] = 8
df.loc[(df['Date'] >= '2021-01-15') & (df['Date'] <= '2021-02-07'), 'Swarm'] = 9
df.loc[(df['Date'] >= '2021-06-01') & (df['Date'] <= '2021-06-30'), 'Swarm'] = 10
df.loc[(df['Date'] >= '2021-07-01'), 'Swarm'] = None


df['Phase'] = 0;
df.loc[(df['Date'] >= '2021-09-11') & (df['Date'] <= '2021-09-30'), 'Phase'] = 1
df.loc[(df['Date'] >= '2021-10-01') & (df['Date'] <= '2021-11-30'), 'Phase'] = 2
df.loc[(df['Date'] >= '2021-12-01') & (df['Date'] <= '2021-12-31'), 'Phase'] = 3
df.loc[(df['Date'] >= '2021-12-31'), 'Phase'] = 4
In [11]:
df.to_csv("../data/lapalma_ign.csv", index=False)