First version of the dataset

README.md

@@ -0,0 +1,98 @@
+---
+language:
+- English
+license:
+- cc-by-4.0
+multilinguality:
+- monolingual
+pretty_name: NOAA Buoy meterological data
+size_categories:
+- 100K<n<1M
+source_datasets:
+- original
+tags: []
+task_categories:
+- feature-extraction
+- tabular-classification
+- time-series-forecasting
+---
+
+# Dataset Card for {{ pretty_name | default("Dataset Name", true) }}
+
+NOAA Buoy Data was downloaded, processed, and cleaned for tasks pertaining to tabular data. The data consists of meteorological measurements. There are two datasets
+
+1. From 1980 through 2022 (denoted with "years" in file names)
+1. From Jan 2023 through end of Sept 2023 (denoted with "2023" in file names)
+
+The original intended use is for anomaly detection in tabular data. 
+
+## Dataset Details
+
+### Dataset Description
+
+This dataset contains weather buoy data to be used in a tabular embedding scenarios. 
+Buoy 42002 was chosen because it had many years of historical data and was still actively collecting information
+
+Here is the buoy's page and its historical data page. 
+https://www.ndbc.noaa.gov/station_page.php?station=42002
+https://www.ndbc.noaa.gov/station_history.php?station=42002
+
+Only standard meteorological data and ocean data was downloaded. Downloaded started at 1980, which is the first full year of collecting wave information.
+
+### Data Fields
+
+{'TSTMP': 'timestamp',
+'#YY': '#yr',
+' MM': 'mo',
+'DD': 'dy',
+'hh': 'hr',
+'mm': 'mn',
+'WDIR': 'degT',
+'WSPD': 'm/s',
+' GST': 'm/s',
+' WVHT': 'm',
+'DPD': 'sec',
+'APD': 'sec',
+'MWD   ': 'degT',
+'PRES': 'hPa',
+' ATMP': 'degC',
+' WTMP': 'degC'
+}
+
+## Dataset Creation
+
+### Curation Rationale
+
+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 
+but kept separate to allow for train and inference. 
+
+### Source Data
+
+#### Initial Data Collection and Normalization
+Data Downloaded on Oct 12 2023
+
+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. 
+
+1. #YY, MM, DD, hh, mm were concatenated to create a timestamp and stored in a new column. 
+1. From 1980 until 2005 there was no recording of minutes. Minutes for those years was set to 00. 
+1. All missing data was set to a blank value rather than an actual number
+1. Remove all rows without wave data from all the data sets ( missing value in WVHT and DPD)
+1. Columns MWD, DEWP, VIS, and TIDE were removed because of consistent missing values
+1. From 2005 -> 2006 Wind direction goes from being called WD to WDIR
+1. From 2006 -> 2007 Header goes from just 1 line with variable names to 2 lines with the second line being units.
+
+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
+
+Since the original purpose of this data was anomoly detection. The two data sets above received further processing. 
+
+1. All data values were converted to Z-scores (file named zscore_2023)
+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 )
+
+## Uses
+
+### Direct Use
+
+Primary use is working with tabular data and embeddings, particularly for anomaly detection
+
+
+

buoy-python/README.md

@@ -0,0 +1,9 @@
+np.mean
+https://numpy.org/doc/stable/reference/generated/numpy.mean.html
+
+np.std
+
+Var Name: #YY  MM DD hh mm WDIR WSPD GST  WVHT   DPD   APD MWD   PRES  ATMP  WTMP  DEWP  VIS  TIDE
+Units:    #yr  mo dy hr mn degT m/s  m/s     m   sec   sec degT   hPa  degC  degC  degC   mi    ft
+
+DEWP, VIS, & TIDE are always missing, so they were removed from the data set

buoy-python/StandardizeAndClean.py

@@ -0,0 +1,168 @@
+import csv
+import time
+from time import strptime
+from datetime import datetime
+from pathlib import Path
+
+# UGLY - the non 2023 functions should be more generic given a certain start location - that way we don't have to repeat
+# logic
+
+# Function for Years
+YEARS_LOCATION = "../orig_downloads/csv"
+LOCATION_2023 = "../orig_downloads/2023/csv"
+
+YEARS_PATH = Path(YEARS_LOCATION)
+YEARS_PATH_2023 = Path(LOCATION_2023)
+
+FINAL_BIG_FILE = "../full_years_remove_flawed_rows.csv"
+FINAL_BIG_FILE_2023 = "../full_2023_remove_flawed_rows.csv"
+
+HEADER = "#YY,MM,DD,hh,mm,WDIR,WSPD,GST,WVHT,DPD,APD,MWD,PRES,ATMP,WTMP,DEWP,VIS,TIDE\n"
+FINAL_HEADER = ["TSTMP", "#YY","MM","DD", "hh","mm","WDIR","WSPD","GST","WVHT","DPD","APD","MWD","PRES","ATMP","WTMP"]
+
+
+# Deal with the difference between files and get them standardized
+def standardize():
+    for read_path in YEARS_PATH.rglob('*.csv'):
+        out_file_name = "fixed_" + read_path.name
+        write_path = str(read_path).replace(read_path.name, out_file_name)
+        with open(read_path, newline='') as read_file, open(write_path, 'w', newline='\n') as write_file:
+            year = read_path.name[6:10]
+            year = int(year)
+            if year <= 2006:
+                # First write the new header line
+                read_file.readline()
+
+                write_file.write(HEADER)
+                for line in read_file:
+                    line = line.strip()
+                    if line[len(line)-1] == ",":
+                        line_array = line[:-1].split(',')
+                    else:
+                        line_array = line.split(',')
+
+                    # pre 1999 we need to make the year 4 digits
+                    if year <= 1998:
+                        line_array[0] = "19" + (line_array[0])
+
+                    # Add tide with a value of 99.00 for all years pre 2000
+                    if year < 2000:
+                        line_array.append('99.0')
+
+                    # Add 0 in for mm pre 2005 (header and values)
+                    if year < 2005:
+                        line_array.insert(4, '0')
+
+                    # Changes are done, write the line
+                    write_file.write(','.join(line_array) + "\n")
+            if year > 2006:
+
+                # Remove second header line from 2007 onwards
+                read_file.readline()
+                read_file.readline()
+
+                # Add the first line back and just write the rest of the lines
+                write_file.write(HEADER)
+                for line in read_file:
+                    line = line.strip()
+                    if line[len(line)-1] == ",":
+                        line = line[0:-1]
+                    write_file.write(line + "\n")
+
+# Now remove the columns we don't want and erase rows with a lot of missing values in columns we care about
+def winnow_down(big_file_name, read_location):
+
+    # need to be become missing data
+    nine9_0 = {"WVHT", "WSPD", "GST", "DPD", "APD"}
+    nine99_0 = {"ATMP", "WTMP"}
+    nine99 = {"WDIR", "MWD"}
+    if_all_missing = {"DPD","APD"}
+    remove_me = {"DEWP", "VIS", "TIDE"}
+
+
+    # Set up the file to write to
+    with open(big_file_name, 'w', newline='') as file:
+        fieldnames = FINAL_HEADER
+        output_csvfile = csv.DictWriter(file, fieldnames=fieldnames)
+
+        output_csvfile.writeheader()
+        for read_path in read_location.rglob('fixed_*.csv'):
+            print(read_path)
+            with open(read_path, newline='') as csv_file:
+                csv_reader = csv.DictReader(csv_file)
+
+                # row is not an ordered dict
+                for row in csv_reader:
+
+                    # Check to see if we are missing key data - if so delete the row and move along
+                    delete_row = 0.0
+                    if row["WSPD"] == "99.0":
+                        delete_row = delete_row + 1.0
+                    if row["WVHT"] == "99.0" or row["WVHT"] == "99.00":
+                        delete_row = delete_row + 1.0
+                    if row["WTMP"] == "999.0":
+                        delete_row = delete_row + 1.0
+                    # if DPD and APD are missing along with any of the above then we remove
+                    for key in if_all_missing:
+                        if row[key] == "99.0" or row[key] == "99.00":
+                            delete_row = delete_row + 0.5
+
+
+                    if delete_row >= 2.0:
+                        # Two strikes you are out and we go on to the next row
+                        continue
+
+                    # Remove observations at least 2 of these columns with null values in wspd (99.0)  wvht (99.0) and wtmp (999.0)
+                    # WD MWD = 999, GST DPD APD = 99.0, PRES = 9999.0, ATMP WTMP = 999.0
+                    # For those left we need to convert these to missing(just a blank)
+                    for key in nine99:
+                        if row[key] == '999':
+                            row[key] = ''
+                    for key in nine9_0:
+                        if row[key] == '99.0' or row[key] == '99.00':
+                            row[key] = ''
+                    for key in nine99_0:
+                        if row[key] == '999.0':
+                            row[key] = ''
+                    if row["PRES"] == '9999.0':
+                        row["PRES"] = ''
+
+                    # remove columns DEMP, VIS, TIDE
+                    for key in remove_me:
+                        del row[key]
+
+                    # Finally we need to convert Y, M, D, m into a timestamp and that will be the key
+                    # Buoy 42002 is in Lousiana, UTC -5
+                    timestamp_string = row["#YY"] + "-" + row["MM"] + "-" + row["DD"] + " " + row["hh"] + ":" + row["mm"] + "-" + "-0500"
+                    row["TSTMP"] = datetime.strptime(timestamp_string, "%Y-%m-%d %H:%M-%z")
+
+                    # Ok we are ready to write a new row to our database
+                    output_csvfile.writerow(row)
+
+# Function for 2023
+def standardize2023():
+    for read_path in YEARS_PATH_2023.rglob('*.csv'):
+        out_file_name = "fixed_" + read_path.name
+        write_path = str(read_path).replace(read_path.name, out_file_name)
+        with open(read_path, newline='') as read_file, open(write_path, 'w', newline='\n') as write_file:
+            # Remove second header line from 2007 onwards
+            read_file.readline()
+            read_file.readline()
+
+            # Add the first line back and just write the rest of the lines
+            write_file.write(HEADER)
+            for line in read_file:
+                line = line.strip()
+                if line[len(line)-1] == ",":
+                    line = line[0:-1]
+                write_file.write(line + "\n")
+
+
+
+if __name__ == '__main__':
+    print("start")
+    #standardize()
+    winnow_down(FINAL_BIG_FILE, YEARS_PATH)
+    #standardize2023()
+    winnow_down(FINAL_BIG_FILE_2023, YEARS_PATH_2023)
+    print("finished")

buoy-python/YearsLessThan2StdDev.py

@@ -0,0 +1,64 @@
+import csv
+from pathlib import Path
+import numpy as np
+import pandas as pd
+
+from collections import defaultdict
+
+FULL_DATA_SET_STRING = "../full_years_remove_flawed_rows.csv"
+FULL_DATA_SET_PATH = Path(FULL_DATA_SET_STRING)
+OUT_DATASET_STRING = "../trimmed_full_years_for_db.parquet"
+OUT_DATASET_PATH = Path(OUT_DATASET_STRING)
+OUT_FULL_DATASET_STRING = "../full_years_remove_flawed.parquet"
+OUT_FULL_DATASET_PATH = Path(OUT_FULL_DATASET_STRING)
+
+NUMERIC_FIELDS = ["WSPD","GST","WVHT","DPD","APD","PRES","ATMP","WTMP"]
+
+
+def load_data(data_path):
+    print("Loading data")
+
+    with open(data_path, newline='') as csv_file:
+
+        loaded_np_data = pd.read_csv(csv_file)
+
+        print("Writing out the full Parquet file")
+        loaded_np_data.to_parquet(OUT_FULL_DATASET_PATH)
+
+        print("Applying Sin() to the two degrees columns")
+        loaded_np_data["WDIR"] = np.sin(np.deg2rad(loaded_np_data["WDIR"]))
+        loaded_np_data["MWD"] = np.sin(np.deg2rad(loaded_np_data["MWD"]))
+
+        print("calculating z-scores")
+        for var in NUMERIC_FIELDS:
+            var_mean = np.mean(loaded_np_data[var])
+            var_std = np.std(loaded_np_data[var])
+
+            var_zscore = (loaded_np_data[var] - var_mean)/var_std
+            loaded_np_data[var] = var_zscore
+
+        print("finding outlier rows")
+
+        # calculate the rows to keep
+        # for each column, is the z-score larger than 2 = loaded_np_data[NUMERIC_FIELDS].le(2)
+        # are there less 2 columns meeting the condition above = keep the row
+        output_np_data = loaded_np_data[loaded_np_data[NUMERIC_FIELDS].gt(2).sum(axis=1).lt(2)]
+
+        print("exporting to parquet")
+        output_np_data.set_index("TSTMP")
+        output_np_data.to_parquet(OUT_DATASET_PATH)
+
+
+if __name__ == '__main__':
+    print("Start")
+
+    # Load data
+    all_data = load_data(FULL_DATA_SET_PATH)
+
+    # Calculate mean and std dev for each non-date column
+    # Going to need to sin(X) for any circular numbers (WDIR & MWD)
+
+
+    # Write out data removing rows
+    # Probably want to write out the sin(X) for any circular numbers
+    print("finished")

buoy-python/ZScore2023.py

@@ -0,0 +1,58 @@
+import csv
+from pathlib import Path
+import numpy as np
+import pandas as pd
+
+from collections import defaultdict
+
+FULL_DATA_SET_STRING = "../full_2023_remove_flawed_rows.csv"
+FULL_DATA_SET_PATH = Path(FULL_DATA_SET_STRING)
+OUT_DATASET_STRING = "../zscore_2023.parquet"
+OUT_DATASET_PATH = Path(OUT_DATASET_STRING)
+OUT_FULL_DATASET_STRING = "../full_2023_remove_flawed.parquet"
+OUT_FULL_DATASET_PATH = Path(OUT_FULL_DATASET_STRING)
+
+NUMERIC_FIELDS = ["WSPD","GST","WVHT","DPD","APD","PRES","ATMP","WTMP"]
+
+
+def load_data(data_path):
+    print("Loading data")
+
+    with open(data_path, newline='') as csv_file:
+
+        loaded_np_data = pd.read_csv(csv_file)
+
+        print("Writing out the full Parquet file")
+        loaded_np_data.to_parquet(OUT_FULL_DATASET_PATH)
+
+        print("Applying Sin() to the two degrees columns")
+        loaded_np_data["WDIR"] = np.sin(np.deg2rad(loaded_np_data["WDIR"]))
+        loaded_np_data["MWD"] = np.sin(np.deg2rad(loaded_np_data["MWD"]))
+
+        print("calculating z-scores")
+        for var in NUMERIC_FIELDS:
+            var_mean = np.mean(loaded_np_data[var])
+            var_std = np.std(loaded_np_data[var])
+
+            var_zscore = (loaded_np_data[var] - var_mean)/var_std
+            loaded_np_data[var] = var_zscore
+
+
+        print("exporting to parquet")
+        loaded_np_data.set_index("TSTMP")
+        loaded_np_data.to_parquet(OUT_DATASET_PATH)
+
+
+if __name__ == '__main__':
+    print("Start")
+
+    # Load data
+    all_data = load_data(FULL_DATA_SET_PATH)
+
+    # Calculate mean and std dev for each non-date column
+    # Going to need to sin(X) for any circular numbers (WDIR & MWD)
+
+
+    # Write out data removing rows
+    # Probably want to write out the sin(X) for any circular numbers
+    print("finished")

buoy-python/requirements.txt

@@ -0,0 +1,8 @@
+pip~=21.3.1
+wheel~=0.37.1
+pytz~=2023.3.post1
+numpy~=1.26.0
+setuptools~=60.2.0
+pandas~=2.1.1
+MarkupSafe~=2.1.1
+python-dateutil~=2.8.2

full_2023_remove_flawed.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ac086d30766c6a26530bf6f067cd4e9ea2c4066cee3eaf851739d3ba805b718
+size 225573

full_2023_remove_flawed_rows.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46f800b7108bd4c9e24b36d9e038188f679553d8c0feb37d2d5e90ad24ee13cb
+size 1173803

full_years_remove_flawed.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b859b371b89f0591319ad85675de2ef1be9a5ecbaaf91887eb7bd0fbe3cce82
+size 4730863

full_years_remove_flawed_rows.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:98043c07483b90f409831eafd758d5f125440118520ec140a02b0c1a95eb3c94
+size 30354345