lukawskikacper
commited on
Commit
·
8be024f
1
Parent(s):
901bf7c
First version of the dataset
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- README.md +98 -0
- buoy-python/README.md +9 -0
- buoy-python/StandardizeAndClean.py +168 -0
- buoy-python/YearsLessThan2StdDev.py +64 -0
- buoy-python/ZScore2023.py +58 -0
- buoy-python/requirements.txt +8 -0
- full_2023_remove_flawed.parquet +3 -0
- full_2023_remove_flawed_rows.csv +3 -0
- full_years_remove_flawed.parquet +3 -0
- full_years_remove_flawed_rows.csv +3 -0
- orig_downloads/2023/42002_Apr.txt +0 -0
- orig_downloads/2023/42002_Aug.txt +0 -0
- orig_downloads/2023/42002_Feb.txt +0 -0
- orig_downloads/2023/42002_Jan.txt +0 -0
- orig_downloads/2023/42002_Jul.txt +0 -0
- orig_downloads/2023/42002_Jun.txt +0 -0
- orig_downloads/2023/42002_Mar.txt +0 -0
- orig_downloads/2023/42002_May.txt +0 -0
- orig_downloads/2023/42002_Sep.txt +0 -0
- orig_downloads/2023/csv/42002_Apr.csv +0 -0
- orig_downloads/2023/csv/42002_Aug.csv +0 -0
- orig_downloads/2023/csv/42002_Feb.csv +0 -0
- orig_downloads/2023/csv/42002_Jan.csv +0 -0
- orig_downloads/2023/csv/42002_Jul.csv +0 -0
- orig_downloads/2023/csv/42002_Jun.csv +0 -0
- orig_downloads/2023/csv/42002_Mar.csv +0 -0
- orig_downloads/2023/csv/42002_May.csv +0 -0
- orig_downloads/2023/csv/42002_Sep.csv +0 -0
- orig_downloads/2023/csv/fixed_42002_Apr.csv +0 -0
- orig_downloads/2023/csv/fixed_42002_Aug.csv +0 -0
- orig_downloads/2023/csv/fixed_42002_Feb.csv +0 -0
- orig_downloads/2023/csv/fixed_42002_Jan.csv +0 -0
- orig_downloads/2023/csv/fixed_42002_Jul.csv +0 -0
- orig_downloads/2023/csv/fixed_42002_Jun.csv +0 -0
- orig_downloads/2023/csv/fixed_42002_Mar.csv +0 -0
- orig_downloads/2023/csv/fixed_42002_May.csv +0 -0
- orig_downloads/2023/csv/fixed_42002_Sep.csv +0 -0
- orig_downloads/42002_1980.txt +0 -0
- orig_downloads/42002_1981.txt +0 -0
- orig_downloads/42002_1982.txt +0 -0
- orig_downloads/42002_1983.txt +0 -0
- orig_downloads/42002_1984.txt +0 -0
- orig_downloads/42002_1985.txt +0 -0
- orig_downloads/42002_1986.txt +0 -0
- orig_downloads/42002_1987.txt +0 -0
- orig_downloads/42002_1988.txt +0 -0
- orig_downloads/42002_1989.txt +0 -0
- orig_downloads/42002_1990.txt +0 -0
- orig_downloads/42002_1991.txt +0 -0
- orig_downloads/42002_1992.txt +0 -0
README.md
ADDED
@@ -0,0 +1,98 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- English
|
4 |
+
license:
|
5 |
+
- cc-by-4.0
|
6 |
+
multilinguality:
|
7 |
+
- monolingual
|
8 |
+
pretty_name: NOAA Buoy meterological data
|
9 |
+
size_categories:
|
10 |
+
- 100K<n<1M
|
11 |
+
source_datasets:
|
12 |
+
- original
|
13 |
+
tags: []
|
14 |
+
task_categories:
|
15 |
+
- feature-extraction
|
16 |
+
- tabular-classification
|
17 |
+
- time-series-forecasting
|
18 |
+
---
|
19 |
+
|
20 |
+
# Dataset Card for {{ pretty_name | default("Dataset Name", true) }}
|
21 |
+
|
22 |
+
NOAA Buoy Data was downloaded, processed, and cleaned for tasks pertaining to tabular data. The data consists of meteorological measurements. There are two datasets
|
23 |
+
|
24 |
+
1. From 1980 through 2022 (denoted with "years" in file names)
|
25 |
+
1. From Jan 2023 through end of Sept 2023 (denoted with "2023" in file names)
|
26 |
+
|
27 |
+
The original intended use is for anomaly detection in tabular data.
|
28 |
+
|
29 |
+
## Dataset Details
|
30 |
+
|
31 |
+
### Dataset Description
|
32 |
+
|
33 |
+
This dataset contains weather buoy data to be used in a tabular embedding scenarios.
|
34 |
+
Buoy 42002 was chosen because it had many years of historical data and was still actively collecting information
|
35 |
+
|
36 |
+
Here is the buoy's page and its historical data page.
|
37 |
+
https://www.ndbc.noaa.gov/station_page.php?station=42002
|
38 |
+
https://www.ndbc.noaa.gov/station_history.php?station=42002
|
39 |
+
|
40 |
+
Only standard meteorological data and ocean data was downloaded. Downloaded started at 1980, which is the first full year of collecting wave information.
|
41 |
+
|
42 |
+
### Data Fields
|
43 |
+
|
44 |
+
{'TSTMP': 'timestamp',
|
45 |
+
'#YY': '#yr',
|
46 |
+
' MM': 'mo',
|
47 |
+
'DD': 'dy',
|
48 |
+
'hh': 'hr',
|
49 |
+
'mm': 'mn',
|
50 |
+
'WDIR': 'degT',
|
51 |
+
'WSPD': 'm/s',
|
52 |
+
' GST': 'm/s',
|
53 |
+
' WVHT': 'm',
|
54 |
+
'DPD': 'sec',
|
55 |
+
'APD': 'sec',
|
56 |
+
'MWD ': 'degT',
|
57 |
+
'PRES': 'hPa',
|
58 |
+
' ATMP': 'degC',
|
59 |
+
' WTMP': 'degC'
|
60 |
+
}
|
61 |
+
|
62 |
+
## Dataset Creation
|
63 |
+
|
64 |
+
### Curation Rationale
|
65 |
+
|
66 |
+
The original data has inconsistent delimiters, different and inappropriate missing data values, and was not harmonized across years. Pre-2023 was edited in the same way as the previous data
|
67 |
+
but kept separate to allow for train and inference.
|
68 |
+
|
69 |
+
### Source Data
|
70 |
+
|
71 |
+
#### Initial Data Collection and Normalization
|
72 |
+
Data Downloaded on Oct 12 2023
|
73 |
+
|
74 |
+
All code used to transform the data can be found in the buoy-python directory. This is NOT production code and the focus was on correct results and minimizing time spent writing cleaning code.
|
75 |
+
|
76 |
+
1. #YY, MM, DD, hh, mm were concatenated to create a timestamp and stored in a new column.
|
77 |
+
1. From 1980 until 2005 there was no recording of minutes. Minutes for those years was set to 00.
|
78 |
+
1. All missing data was set to a blank value rather than an actual number
|
79 |
+
1. Remove all rows without wave data from all the data sets ( missing value in WVHT and DPD)
|
80 |
+
1. Columns MWD, DEWP, VIS, and TIDE were removed because of consistent missing values
|
81 |
+
1. From 2005 -> 2006 Wind direction goes from being called WD to WDIR
|
82 |
+
1. From 2006 -> 2007 Header goes from just 1 line with variable names to 2 lines with the second line being units.
|
83 |
+
|
84 |
+
These steps were used to create full_2023_remove_flawed_rows, the 2023 months, and full_years_remove_flawed_rows the previous data going back to 1980
|
85 |
+
|
86 |
+
Since the original purpose of this data was anomoly detection. The two data sets above received further processing.
|
87 |
+
|
88 |
+
1. All data values were converted to Z-scores (file named zscore_2023)
|
89 |
+
1. For 1980 - 2022, all rows with 2 or more fields with Z-scores > 2 were removed from the dataset (file named trimmed_zscores_years )
|
90 |
+
|
91 |
+
## Uses
|
92 |
+
|
93 |
+
### Direct Use
|
94 |
+
|
95 |
+
Primary use is working with tabular data and embeddings, particularly for anomaly detection
|
96 |
+
|
97 |
+
|
98 |
+
|
buoy-python/README.md
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
np.mean
|
2 |
+
https://numpy.org/doc/stable/reference/generated/numpy.mean.html
|
3 |
+
|
4 |
+
np.std
|
5 |
+
|
6 |
+
Var Name: #YY MM DD hh mm WDIR WSPD GST WVHT DPD APD MWD PRES ATMP WTMP DEWP VIS TIDE
|
7 |
+
Units: #yr mo dy hr mn degT m/s m/s m sec sec degT hPa degC degC degC mi ft
|
8 |
+
|
9 |
+
DEWP, VIS, & TIDE are always missing, so they were removed from the data set
|
buoy-python/StandardizeAndClean.py
ADDED
@@ -0,0 +1,168 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import csv
|
2 |
+
import time
|
3 |
+
from time import strptime
|
4 |
+
from datetime import datetime
|
5 |
+
from pathlib import Path
|
6 |
+
|
7 |
+
# UGLY - the non 2023 functions should be more generic given a certain start location - that way we don't have to repeat
|
8 |
+
# logic
|
9 |
+
|
10 |
+
# Function for Years
|
11 |
+
YEARS_LOCATION = "../orig_downloads/csv"
|
12 |
+
LOCATION_2023 = "../orig_downloads/2023/csv"
|
13 |
+
|
14 |
+
YEARS_PATH = Path(YEARS_LOCATION)
|
15 |
+
YEARS_PATH_2023 = Path(LOCATION_2023)
|
16 |
+
|
17 |
+
FINAL_BIG_FILE = "../full_years_remove_flawed_rows.csv"
|
18 |
+
FINAL_BIG_FILE_2023 = "../full_2023_remove_flawed_rows.csv"
|
19 |
+
|
20 |
+
HEADER = "#YY,MM,DD,hh,mm,WDIR,WSPD,GST,WVHT,DPD,APD,MWD,PRES,ATMP,WTMP,DEWP,VIS,TIDE\n"
|
21 |
+
FINAL_HEADER = ["TSTMP", "#YY","MM","DD", "hh","mm","WDIR","WSPD","GST","WVHT","DPD","APD","MWD","PRES","ATMP","WTMP"]
|
22 |
+
|
23 |
+
|
24 |
+
# Deal with the difference between files and get them standardized
|
25 |
+
def standardize():
|
26 |
+
for read_path in YEARS_PATH.rglob('*.csv'):
|
27 |
+
out_file_name = "fixed_" + read_path.name
|
28 |
+
write_path = str(read_path).replace(read_path.name, out_file_name)
|
29 |
+
with open(read_path, newline='') as read_file, open(write_path, 'w', newline='\n') as write_file:
|
30 |
+
year = read_path.name[6:10]
|
31 |
+
year = int(year)
|
32 |
+
if year <= 2006:
|
33 |
+
# First write the new header line
|
34 |
+
read_file.readline()
|
35 |
+
|
36 |
+
write_file.write(HEADER)
|
37 |
+
for line in read_file:
|
38 |
+
line = line.strip()
|
39 |
+
if line[len(line)-1] == ",":
|
40 |
+
line_array = line[:-1].split(',')
|
41 |
+
else:
|
42 |
+
line_array = line.split(',')
|
43 |
+
|
44 |
+
# pre 1999 we need to make the year 4 digits
|
45 |
+
if year <= 1998:
|
46 |
+
line_array[0] = "19" + (line_array[0])
|
47 |
+
|
48 |
+
# Add tide with a value of 99.00 for all years pre 2000
|
49 |
+
if year < 2000:
|
50 |
+
line_array.append('99.0')
|
51 |
+
|
52 |
+
# Add 0 in for mm pre 2005 (header and values)
|
53 |
+
if year < 2005:
|
54 |
+
line_array.insert(4, '0')
|
55 |
+
|
56 |
+
# Changes are done, write the line
|
57 |
+
write_file.write(','.join(line_array) + "\n")
|
58 |
+
if year > 2006:
|
59 |
+
|
60 |
+
# Remove second header line from 2007 onwards
|
61 |
+
read_file.readline()
|
62 |
+
read_file.readline()
|
63 |
+
|
64 |
+
# Add the first line back and just write the rest of the lines
|
65 |
+
write_file.write(HEADER)
|
66 |
+
for line in read_file:
|
67 |
+
line = line.strip()
|
68 |
+
if line[len(line)-1] == ",":
|
69 |
+
line = line[0:-1]
|
70 |
+
write_file.write(line + "\n")
|
71 |
+
|
72 |
+
# Now remove the columns we don't want and erase rows with a lot of missing values in columns we care about
|
73 |
+
def winnow_down(big_file_name, read_location):
|
74 |
+
|
75 |
+
# need to be become missing data
|
76 |
+
nine9_0 = {"WVHT", "WSPD", "GST", "DPD", "APD"}
|
77 |
+
nine99_0 = {"ATMP", "WTMP"}
|
78 |
+
nine99 = {"WDIR", "MWD"}
|
79 |
+
if_all_missing = {"DPD","APD"}
|
80 |
+
remove_me = {"DEWP", "VIS", "TIDE"}
|
81 |
+
|
82 |
+
|
83 |
+
# Set up the file to write to
|
84 |
+
with open(big_file_name, 'w', newline='') as file:
|
85 |
+
fieldnames = FINAL_HEADER
|
86 |
+
output_csvfile = csv.DictWriter(file, fieldnames=fieldnames)
|
87 |
+
|
88 |
+
output_csvfile.writeheader()
|
89 |
+
for read_path in read_location.rglob('fixed_*.csv'):
|
90 |
+
print(read_path)
|
91 |
+
with open(read_path, newline='') as csv_file:
|
92 |
+
csv_reader = csv.DictReader(csv_file)
|
93 |
+
|
94 |
+
# row is not an ordered dict
|
95 |
+
for row in csv_reader:
|
96 |
+
|
97 |
+
# Check to see if we are missing key data - if so delete the row and move along
|
98 |
+
delete_row = 0.0
|
99 |
+
if row["WSPD"] == "99.0":
|
100 |
+
delete_row = delete_row + 1.0
|
101 |
+
if row["WVHT"] == "99.0" or row["WVHT"] == "99.00":
|
102 |
+
delete_row = delete_row + 1.0
|
103 |
+
if row["WTMP"] == "999.0":
|
104 |
+
delete_row = delete_row + 1.0
|
105 |
+
# if DPD and APD are missing along with any of the above then we remove
|
106 |
+
for key in if_all_missing:
|
107 |
+
if row[key] == "99.0" or row[key] == "99.00":
|
108 |
+
delete_row = delete_row + 0.5
|
109 |
+
|
110 |
+
|
111 |
+
if delete_row >= 2.0:
|
112 |
+
# Two strikes you are out and we go on to the next row
|
113 |
+
continue
|
114 |
+
|
115 |
+
# Remove observations at least 2 of these columns with null values in wspd (99.0) wvht (99.0) and wtmp (999.0)
|
116 |
+
# WD MWD = 999, GST DPD APD = 99.0, PRES = 9999.0, ATMP WTMP = 999.0
|
117 |
+
# For those left we need to convert these to missing(just a blank)
|
118 |
+
for key in nine99:
|
119 |
+
if row[key] == '999':
|
120 |
+
row[key] = ''
|
121 |
+
for key in nine9_0:
|
122 |
+
if row[key] == '99.0' or row[key] == '99.00':
|
123 |
+
row[key] = ''
|
124 |
+
for key in nine99_0:
|
125 |
+
if row[key] == '999.0':
|
126 |
+
row[key] = ''
|
127 |
+
if row["PRES"] == '9999.0':
|
128 |
+
row["PRES"] = ''
|
129 |
+
|
130 |
+
# remove columns DEMP, VIS, TIDE
|
131 |
+
for key in remove_me:
|
132 |
+
del row[key]
|
133 |
+
|
134 |
+
# Finally we need to convert Y, M, D, m into a timestamp and that will be the key
|
135 |
+
# Buoy 42002 is in Lousiana, UTC -5
|
136 |
+
timestamp_string = row["#YY"] + "-" + row["MM"] + "-" + row["DD"] + " " + row["hh"] + ":" + row["mm"] + "-" + "-0500"
|
137 |
+
row["TSTMP"] = datetime.strptime(timestamp_string, "%Y-%m-%d %H:%M-%z")
|
138 |
+
|
139 |
+
# Ok we are ready to write a new row to our database
|
140 |
+
output_csvfile.writerow(row)
|
141 |
+
|
142 |
+
# Function for 2023
|
143 |
+
def standardize2023():
|
144 |
+
for read_path in YEARS_PATH_2023.rglob('*.csv'):
|
145 |
+
out_file_name = "fixed_" + read_path.name
|
146 |
+
write_path = str(read_path).replace(read_path.name, out_file_name)
|
147 |
+
with open(read_path, newline='') as read_file, open(write_path, 'w', newline='\n') as write_file:
|
148 |
+
# Remove second header line from 2007 onwards
|
149 |
+
read_file.readline()
|
150 |
+
read_file.readline()
|
151 |
+
|
152 |
+
# Add the first line back and just write the rest of the lines
|
153 |
+
write_file.write(HEADER)
|
154 |
+
for line in read_file:
|
155 |
+
line = line.strip()
|
156 |
+
if line[len(line)-1] == ",":
|
157 |
+
line = line[0:-1]
|
158 |
+
write_file.write(line + "\n")
|
159 |
+
|
160 |
+
|
161 |
+
|
162 |
+
if __name__ == '__main__':
|
163 |
+
print("start")
|
164 |
+
#standardize()
|
165 |
+
winnow_down(FINAL_BIG_FILE, YEARS_PATH)
|
166 |
+
#standardize2023()
|
167 |
+
winnow_down(FINAL_BIG_FILE_2023, YEARS_PATH_2023)
|
168 |
+
print("finished")
|
buoy-python/YearsLessThan2StdDev.py
ADDED
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import csv
|
2 |
+
from pathlib import Path
|
3 |
+
import numpy as np
|
4 |
+
import pandas as pd
|
5 |
+
|
6 |
+
from collections import defaultdict
|
7 |
+
|
8 |
+
FULL_DATA_SET_STRING = "../full_years_remove_flawed_rows.csv"
|
9 |
+
FULL_DATA_SET_PATH = Path(FULL_DATA_SET_STRING)
|
10 |
+
OUT_DATASET_STRING = "../trimmed_full_years_for_db.parquet"
|
11 |
+
OUT_DATASET_PATH = Path(OUT_DATASET_STRING)
|
12 |
+
OUT_FULL_DATASET_STRING = "../full_years_remove_flawed.parquet"
|
13 |
+
OUT_FULL_DATASET_PATH = Path(OUT_FULL_DATASET_STRING)
|
14 |
+
|
15 |
+
NUMERIC_FIELDS = ["WSPD","GST","WVHT","DPD","APD","PRES","ATMP","WTMP"]
|
16 |
+
|
17 |
+
|
18 |
+
def load_data(data_path):
|
19 |
+
print("Loading data")
|
20 |
+
|
21 |
+
with open(data_path, newline='') as csv_file:
|
22 |
+
|
23 |
+
loaded_np_data = pd.read_csv(csv_file)
|
24 |
+
|
25 |
+
print("Writing out the full Parquet file")
|
26 |
+
loaded_np_data.to_parquet(OUT_FULL_DATASET_PATH)
|
27 |
+
|
28 |
+
print("Applying Sin() to the two degrees columns")
|
29 |
+
loaded_np_data["WDIR"] = np.sin(np.deg2rad(loaded_np_data["WDIR"]))
|
30 |
+
loaded_np_data["MWD"] = np.sin(np.deg2rad(loaded_np_data["MWD"]))
|
31 |
+
|
32 |
+
print("calculating z-scores")
|
33 |
+
for var in NUMERIC_FIELDS:
|
34 |
+
var_mean = np.mean(loaded_np_data[var])
|
35 |
+
var_std = np.std(loaded_np_data[var])
|
36 |
+
|
37 |
+
var_zscore = (loaded_np_data[var] - var_mean)/var_std
|
38 |
+
loaded_np_data[var] = var_zscore
|
39 |
+
|
40 |
+
print("finding outlier rows")
|
41 |
+
|
42 |
+
# calculate the rows to keep
|
43 |
+
# for each column, is the z-score larger than 2 = loaded_np_data[NUMERIC_FIELDS].le(2)
|
44 |
+
# are there less 2 columns meeting the condition above = keep the row
|
45 |
+
output_np_data = loaded_np_data[loaded_np_data[NUMERIC_FIELDS].gt(2).sum(axis=1).lt(2)]
|
46 |
+
|
47 |
+
print("exporting to parquet")
|
48 |
+
output_np_data.set_index("TSTMP")
|
49 |
+
output_np_data.to_parquet(OUT_DATASET_PATH)
|
50 |
+
|
51 |
+
|
52 |
+
if __name__ == '__main__':
|
53 |
+
print("Start")
|
54 |
+
|
55 |
+
# Load data
|
56 |
+
all_data = load_data(FULL_DATA_SET_PATH)
|
57 |
+
|
58 |
+
# Calculate mean and std dev for each non-date column
|
59 |
+
# Going to need to sin(X) for any circular numbers (WDIR & MWD)
|
60 |
+
|
61 |
+
|
62 |
+
# Write out data removing rows
|
63 |
+
# Probably want to write out the sin(X) for any circular numbers
|
64 |
+
print("finished")
|
buoy-python/ZScore2023.py
ADDED
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import csv
|
2 |
+
from pathlib import Path
|
3 |
+
import numpy as np
|
4 |
+
import pandas as pd
|
5 |
+
|
6 |
+
from collections import defaultdict
|
7 |
+
|
8 |
+
FULL_DATA_SET_STRING = "../full_2023_remove_flawed_rows.csv"
|
9 |
+
FULL_DATA_SET_PATH = Path(FULL_DATA_SET_STRING)
|
10 |
+
OUT_DATASET_STRING = "../zscore_2023.parquet"
|
11 |
+
OUT_DATASET_PATH = Path(OUT_DATASET_STRING)
|
12 |
+
OUT_FULL_DATASET_STRING = "../full_2023_remove_flawed.parquet"
|
13 |
+
OUT_FULL_DATASET_PATH = Path(OUT_FULL_DATASET_STRING)
|
14 |
+
|
15 |
+
NUMERIC_FIELDS = ["WSPD","GST","WVHT","DPD","APD","PRES","ATMP","WTMP"]
|
16 |
+
|
17 |
+
|
18 |
+
def load_data(data_path):
|
19 |
+
print("Loading data")
|
20 |
+
|
21 |
+
with open(data_path, newline='') as csv_file:
|
22 |
+
|
23 |
+
loaded_np_data = pd.read_csv(csv_file)
|
24 |
+
|
25 |
+
print("Writing out the full Parquet file")
|
26 |
+
loaded_np_data.to_parquet(OUT_FULL_DATASET_PATH)
|
27 |
+
|
28 |
+
print("Applying Sin() to the two degrees columns")
|
29 |
+
loaded_np_data["WDIR"] = np.sin(np.deg2rad(loaded_np_data["WDIR"]))
|
30 |
+
loaded_np_data["MWD"] = np.sin(np.deg2rad(loaded_np_data["MWD"]))
|
31 |
+
|
32 |
+
print("calculating z-scores")
|
33 |
+
for var in NUMERIC_FIELDS:
|
34 |
+
var_mean = np.mean(loaded_np_data[var])
|
35 |
+
var_std = np.std(loaded_np_data[var])
|
36 |
+
|
37 |
+
var_zscore = (loaded_np_data[var] - var_mean)/var_std
|
38 |
+
loaded_np_data[var] = var_zscore
|
39 |
+
|
40 |
+
|
41 |
+
print("exporting to parquet")
|
42 |
+
loaded_np_data.set_index("TSTMP")
|
43 |
+
loaded_np_data.to_parquet(OUT_DATASET_PATH)
|
44 |
+
|
45 |
+
|
46 |
+
if __name__ == '__main__':
|
47 |
+
print("Start")
|
48 |
+
|
49 |
+
# Load data
|
50 |
+
all_data = load_data(FULL_DATA_SET_PATH)
|
51 |
+
|
52 |
+
# Calculate mean and std dev for each non-date column
|
53 |
+
# Going to need to sin(X) for any circular numbers (WDIR & MWD)
|
54 |
+
|
55 |
+
|
56 |
+
# Write out data removing rows
|
57 |
+
# Probably want to write out the sin(X) for any circular numbers
|
58 |
+
print("finished")
|
buoy-python/requirements.txt
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
pip~=21.3.1
|
2 |
+
wheel~=0.37.1
|
3 |
+
pytz~=2023.3.post1
|
4 |
+
numpy~=1.26.0
|
5 |
+
setuptools~=60.2.0
|
6 |
+
pandas~=2.1.1
|
7 |
+
MarkupSafe~=2.1.1
|
8 |
+
python-dateutil~=2.8.2
|
full_2023_remove_flawed.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9ac086d30766c6a26530bf6f067cd4e9ea2c4066cee3eaf851739d3ba805b718
|
3 |
+
size 225573
|
full_2023_remove_flawed_rows.csv
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:46f800b7108bd4c9e24b36d9e038188f679553d8c0feb37d2d5e90ad24ee13cb
|
3 |
+
size 1173803
|
full_years_remove_flawed.parquet
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4b859b371b89f0591319ad85675de2ef1be9a5ecbaaf91887eb7bd0fbe3cce82
|
3 |
+
size 4730863
|
full_years_remove_flawed_rows.csv
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:98043c07483b90f409831eafd758d5f125440118520ec140a02b0c1a95eb3c94
|
3 |
+
size 30354345
|
orig_downloads/2023/42002_Apr.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/42002_Aug.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/42002_Feb.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/42002_Jan.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/42002_Jul.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/42002_Jun.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/42002_Mar.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/42002_May.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/42002_Sep.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/42002_Apr.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/42002_Aug.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/42002_Feb.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/42002_Jan.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/42002_Jul.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/42002_Jun.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/42002_Mar.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/42002_May.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/42002_Sep.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/fixed_42002_Apr.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/fixed_42002_Aug.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/fixed_42002_Feb.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/fixed_42002_Jan.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/fixed_42002_Jul.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/fixed_42002_Jun.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/fixed_42002_Mar.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/fixed_42002_May.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/2023/csv/fixed_42002_Sep.csv
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/42002_1980.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/42002_1981.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/42002_1982.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/42002_1983.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/42002_1984.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/42002_1985.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/42002_1986.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/42002_1987.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/42002_1988.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/42002_1989.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/42002_1990.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/42002_1991.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
orig_downloads/42002_1992.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|