-
Notifications
You must be signed in to change notification settings - Fork 0
/
process_data.py
196 lines (156 loc) · 6.63 KB
/
process_data.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
# This script processes all raw datasets into a more compact form, removing all unnecessary features.
# Also contains functions to create new time-series datasets.
# EPA data: https://aqs.epa.gov/aqsweb/airdata/download_files.html#AQI
# Meswest data: https://mesowest.utah.edu/
# NOAA data: https://www.ncdc.noaa.gov/cdo-web/search
# EPA data is originally in the form [State, County, State Code, County Code, Date, AQI, Category, Defining Parameter, Defining Site, # of Sites]
# We want to focus on Salt Lake county so we will remove all other data. Also need to remove unneeded features.
# Meso data units:
# Temperature - C
# Wind_X, Wind_Y - m/s
# Humidity - %
# Pressure - Pascals
import csv
import math
import dateutil.parser as parser
from datetime import timedelta
from os import listdir
from operator import add
# Checks if the input exists, if not returns 0, otherwise converts it to a float
def verify_input(data):
if data == '':
return 0.0
else:
return float(data)
# Processes a AQI data file from the EPA
def process_epa_data(file_name):
data = {}
with open(file_name) as csv_file:
csv_reader = csv.reader(csv_file, delimiter=',')
# Look for rows with Utah, Salt Lake
for row in csv_reader:
if row[0] == 'Utah' and row[1] == 'Salt Lake':
data[row[4]] = [row[5], row[6], row[7]]
return data
# Processes all files in raw_epa_data folder, creating a single epa_data csv file
def create_epa_dataset():
with open('epa_data.csv', 'w', newline='') as csvfile:
writer = csv.writer(csvfile, delimiter=',')
writer.writerow(['Date', 'AQI', 'Category', 'Parameter']) # Adds a header
# Iterate over all datasets
for f in listdir('raw_epa_data'):
processed_data = process_epa_data('raw_epa_data/' + f)
for date, data in processed_data.items():
writer.writerow([date] + data)
# Processes weather data from MesoWest
# Some of the data is useless, while some of it is recorded every hour
# Thus we need to process the data into a more useful format, such as avg. temp for the day, avg. wind speed, etc.
# The first index will be the date in the format 'YYYY-MM-DD' to match the EPA data
# Current output is [Date] = [Temp (C), Wind (m/s), Humidity (%), Pressure (P)]
def process_meso_data(file_name):
processed_data = {}
with open(file_name) as csv_file:
csv_reader = csv.reader(csv_file, delimiter=',')
# Get the CSV headers and units
header = next(csv_reader)
units = next(csv_reader)
# Determine indices to access
temp_index = header.index('air_temp_set_1')
humidity_index = header.index('relative_humidity_set_1')
speed_index = header.index('wind_speed_set_1')
angle_index = header.index('wind_direction_set_1')
pressure_index = header.index('pressure_set_1d')
# Initialize stuff on the first line
# We need to group data by the day, so keep a running count of all daily averages we would like to track
first_line = next(csv_reader)
current_date = first_line[1].split('T', 2)[0]
avg_temp = verify_input(first_line[temp_index])
avg_humidity = verify_input(first_line[humidity_index])
speed = verify_input(first_line[speed_index])
angle = verify_input(first_line[angle_index])
avg_wind = [speed * math.cos(math.radians(angle)), speed * math.sin(math.radians(angle))]
avg_pressure = verify_input(first_line[pressure_index])
count = 1.0
for row in csv_reader:
date, time = row[1].split('T', 2) # Get the date from the timestamp
# End of sequence, store averages
if date != current_date:
processed_data[current_date] = [avg_temp/count, avg_wind[0]/count, avg_wind[1]/count, avg_humidity/count, avg_pressure/count]
avg_temp = 0.0
avg_wind = [0.0, 0.0]
avg_humidity = 0.0
avg_pressure = 0.0
count = 0.0
current_date = date
# Update averages
avg_temp += verify_input(row[temp_index])
avg_humidity += verify_input(row[humidity_index])
speed = verify_input(row[speed_index])
angle = verify_input(row[angle_index])
avg_wind[0] += speed * math.cos(math.radians(angle))
avg_wind[1] += speed * math.sin(math.radians(angle))
avg_pressure += verify_input(row[pressure_index])
count += 1
# Dump the last value
processed_data[current_date] = [avg_temp/count, avg_wind[0]/count, avg_wind[1]/count, avg_humidity/count, avg_pressure/count]
return processed_data
# Processes all raw data from MesoWest into a single CSV file
# Raw data that shares the same date will be merged and averaged
def create_meso_dataset():
datasets = []
processed_data = {}
# First collect all the datasets to be merged
for f in listdir('raw_meso_data'):
datasets.append(process_meso_data('raw_meso_data/' + f))
# Now combine them
merge_count = {}
for dataset in datasets:
for date, data in dataset.items():
if date in merge_count:
processed_data[date] = list(map(add, processed_data[date], data))
merge_count[date] += 1
else:
merge_count[date] = 1
processed_data[date] = data
# Divide each element by the mergecount to average them
for date, data in processed_data.items():
processed_data[date] = [x / merge_count[date] for x in data]
# Write the merged data to a csv file
with open('meso_data.csv', 'w', newline='') as csvfile:
writer = csv.writer(csvfile, delimiter=',')
writer.writerow(['Date', 'Temperature', 'Wind_X', 'Wind_Y', 'Humidity', 'Pressure']) # Adds a header
for date, data in processed_data.items():
writer.writerow([date] + data)
return processed_data
# Loads a dataset, returning a list containing the headers and a dictionary containing the data
def load_dataset(file_name):
header = []
data = {}
with open(file_name) as csv_file:
csv_reader = csv.reader(csv_file, delimiter=',')
header = next(csv_reader)
header.pop(0) # Gets read of the date part
for row in csv_reader:
data[row[0]] = row[1:]
return header, data
# Creates a time series model from a dataset
# Assumes last index of dataset is the label
def time_series(dataset):
new_dataset = []
# For every element except the last, relabel the example using the next example's label
for i in range(len(dataset) - 1):
new_dataset.append(dataset[i] + [dataset[i+1][-1]])
return new_dataset
# Converts a categorical label into a set of classes
def convert_class(dataset):
converted_dataset = list(dataset)
classes = {}
class_count = 0
for i in range(len(dataset)):
class_label = dataset[i][-1]
# Found a new label
if class_label not in classes:
class_count += 1
classes[class_label] = class_count
converted_dataset[i][-1] = classes[class_label]
return classes, converted_dataset