Refactor: Fixed-distance statistics

air_traffic/trajectory.py

@@ -0,0 +1,56 @@
+from geopy import distance
+
+
+def find_first_landed(df):
+
+    # Get through the inital take-off stage
+    n = 0
+    while df["altitude"].iloc[n] == 0:
+        n += 1
+        # Also possible that the whole flight record is already landed
+        if n == len(df):
+            n = 0
+            break
+
+    # Locate the last part of the trajectory
+    df = df.iloc[n:]
+    df = df.loc[df["altitude"] == 0]
+
+    # The first row that has altitude zero is when the flight landed
+    return df.iloc[0]
+
+
+def find_first_in_range(df, d_min, d_max):
+
+    n = 0
+    first = df.iloc[0]
+    d = distance_hkia(first)
+
+    while d > d_max:
+        n += 1
+        first = df.iloc[n]
+        d = distance_hkia(first)
+
+    # Return None if there are no point within the ring
+    if d < d_min:
+        return None
+    else:
+        return first
+
+
+def distance(row1, row2):
+
+    point1 = (row1["latitude"], row1["longitude"])
+    point2 = (row2["latitude"], row2["longitude"])
+
+    d = distance.distance(point1, point2).km
+    return d
+
+
+def distance_hkia(row):
+
+    hkia = (22.308046, 113.918480)
+    point = (row["latitude"], row["longitude"])
+
+    d = distance.distance(point, hkia).km
+    return d

pipeline/3_stat_fixed_distance.py

@@ -0,0 +1,90 @@
+import pandas as pd
+import os
+import numpy as np
+
+from air_traffic.trajectory import *
+
+
+d_min = 145
+d_max = 165
+
+datadir = "../data/cleaned"
+savedir = "../data/results"
+savename = f"stat_fixed_distance_{d_min}-{d_max}.csv"
+
+# --------------------------------------------------------------------------- #
+data = {"date": [],             # Date
+        "callsign": [],         # Callsign
+        "lat_i": [],            # Entry lat
+        "lon_i": [],            # Entry lon
+        "lat_f": [],            # Final lat
+        "lon_f": [],            # Final lon
+        "t_i": [],              # Entry timestamp
+        "t_f": [],              # Final timestamp
+        "r_i_km": [],           # Entry distance from HKIA in km
+        "delta_r_km": [],       # Distance between entry and final point in km
+        "delta_t_sec": []}      # Time difference in second
+
+
+for subdir, dirs, files in os.walk(datadir):
+
+    # Process in sorted order for easy tracking
+    dirs.sort()
+    files.sort()
+
+    for file in files:
+        fname = os.path.join(subdir, file)
+
+        if fname[-4:] != ".csv":
+            continue
+        else:
+            print(fname)
+
+        df = pd.read_csv(fname, header=0)
+        df = df.loc[(df["latitude"] > 19) & (df["latitude"] < 25.5) &
+                    (df["longitude"] > 111) & (df["longitude"] < 117.5)]
+
+        # Skip if there is no useable data
+        if len(df) == 0:
+            continue
+
+        # Find the first landing point
+        # Get through the inital take-off stage
+        last = find_first_landed(df)
+
+        # Check this later
+        # # Skip flights that not landed in the square
+        # if (np.abs(last["latitude"] - hkia[0]) > 0.25) and \
+        #    (np.abs(last["longitude"] - hkia[1]) > 0.25):
+        #     continue
+
+        # Find the first point in the fixed ring
+        first = find_first_in_range(df, d_min, d_max)
+        if first is None:
+            continue
+
+        # Get the date at HK time
+        date = pd.Timestamp(first["time"] + 8*3600, unit="s").strftime("%Y-%m-%d")
+
+        # Identifiers
+        data["date"].append(date)
+        data["callsign"].append(file[:-4])
+
+        last_point = (last["latitude"], last["longitude"])
+
+        # Positional data
+        data["lat_i"].append(first["latitude"])
+        data["lon_i"].append(first["longitude"])
+        data["lat_f"].append(last["latitude"])
+        data["lon_f"].append(last["longitude"])
+        data["delta_r_km"].append(distance(first, last))
+        data["r_i_km"].append(distance_hkia(first))
+
+        # Compute time difference
+        data["t_i"].append(first["time"])
+        data["t_f"].append(last["time"])
+        data["delta_t_sec"].append(last["time"] - first["time"])
+
+
+df_master = pd.DataFrame(data=data)
+df_master.to_csv(f"{savedir}/{savename}", index=False)