First commit for basic analysis code.

apc_data_analysis.py

@@ -0,0 +1,23 @@
+import numpy as np
+import matplotlib.pyplot as plt
+import os
+from apc_plot import APCPlotMixin
+from apc_data_process import APCDataProcessMixin
+
+class APCDataAnalysis(APCPlotMixin, APCDataProcessMixin):
+
+    def __init__(self, load_data=True, clean_data=True):
+        """
+        General initializer
+        """
+        # annoying hard coded paths
+        self.root_dir = os.path.join('/Users/edwardyoung/Google Drive/',
+            'CodeForSF/OpenTransit/OT Raw Data (not in shared drive)')
+        self.clean_data_dir = os.path.join('/Users/edwardyoung/Google Drive/',
+            'CodeForSF/OpenTransit/clean_data')
+
+        if load_data:
+            self.load_data()
+
+        if clean_data:
+            self.clean_data()

apc_data_process.py

@@ -0,0 +1,50 @@
+import numpy as np
+import os
+
+class APCDataProcessMixin:
+
+    def load_data(self, sort_by_date=True):
+        """
+        Load the raw npy files from disk. These have already been extracted
+        from the csv file. This preprocessing makes the data reading much
+        faster.
+        """
+        self.data = {}
+        self.data['OPEN_DATE_TIME'] = np.load(os.path.join(self.clean_data_dir, 
+            'OPEN_DATE_TIME.npy'))
+        self.data['CLOSE_DATE_TIME'] = np.load(os.path.join(self.clean_data_dir, 
+            'CLOSE_DATE_TIME.npy'))
+        self.data['VEH_LAT'] = np.load(os.path.join(self.clean_data_dir,
+            'VEH_LAT.npy'))
+        self.data['VEH_LONG'] = np.load(os.path.join(self.clean_data_dir,
+            'VEH_LONG.npy'))
+        self.data['ONS'] = np.load(os.path.join(self.clean_data_dir, 
+            'ONS.npy'))
+        self.data['OFFS'] = np.load(os.path.join(self.clean_data_dir,
+            'OFFS.npy'))
+        self.data['RUN_ID'] = np.load(os.path.join(self.clean_data_dir, 
+            'RUN_ID.npy'))
+        self.data['ROUTE_ID'] = np.load(os.path.join(self.clean_data_dir, 
+            'ROUTE_ID.npy'))
+
+        if sort_by_date:
+            # Sort by open time
+            idx = np.argsort(self.data['OPEN_DATE_TIME'])
+            for k in self.data.keys():
+                self.data[k] = self.data[k][idx]
+
+
+    def clean_data(self):
+        """
+        Generic function for cleaning data. Currently only cleans out 
+        lat=0/long=0. 
+        """
+        # Mask files with lat = 0 or lon = 0
+        mask = np.where(np.logical_or(self.data['VEH_LAT']==0, 
+            self.data['VEH_LONG']==0))
+
+        for k in self.data.keys():
+            if isinstance(self.data[k][0], float):
+                self.data[k][mask] = np.nan
+
+

apc_plot.py

@@ -0,0 +1,54 @@
+import numpy as np
+import matplotlib.pyplot as plt
+
+class APCPlotMixin:
+
+    def plot_lat_lon(self, start_date, end_date, run_id=None, route_id=None):
+        """
+        Args:
+        -----
+        start_date (datetime) : The beginning of time to plot
+        end_date (datetime) : The end of time to plot
+
+        Opt Args:
+        ---------
+        run_id (int) : The ID of the run.
+        route_id (int) : The route number. What do these 2 things really mean??
+        """
+        # Find indicies of data points between start and end dates
+        idx = np.logical_and(self.data['OPEN_DATE_TIME'] > start_date, 
+            self.data['OPEN_DATE_TIME'] < end_date)
+        if run_id is not None:
+            idx = np.logical_and(idx, self.data['RUN_ID'] == run_id)
+        if route_id is not None:
+            idx = np.logical_and(idx, self.data['ROUTE_ID'] == route_id)
+
+
+        print(f'There are {np.sum(idx)} elements')
+
+        fig, ax = plt.subplots(4, sharex=True, figsize=(5,8))
+        ax[0].plot(self.data['OPEN_DATE_TIME'][idx], self.data['VEH_LAT'][idx])
+        ax[1].plot(self.data['OPEN_DATE_TIME'][idx], self.data['VEH_LONG'][idx])
+        ax[2].plot(self.data['OPEN_DATE_TIME'][idx], self.data['ONS'][idx], 
+            label='On')
+        ax[2].plot(self.data['OPEN_DATE_TIME'][idx], self.data['OFFS'][idx], 
+            label='Off')
+        ax[3].plot(self.data['OPEN_DATE_TIME'][idx], 
+            np.cumsum(self.data['ONS'][idx])-np.cumsum(self.data['OFFS'][idx]))
+
+        ax[2].legend(loc='upper right')
+        ax[0].set_ylabel('Lat')
+        ax[1].set_ylabel('Lon')
+        ax[2].set_ylabel('On/Off #')
+        ax[3].set_ylabel('Total #')
+        ax[3].set_xlabel('Time')
+
+        for tick in ax[3].get_xticklabels():
+            tick.set_rotation(90)
+
+        plt.tight_layout()
+
+        plt.figure()
+        plt.plot(self.data['VEH_LAT'][idx],self.data['VEH_LONG'][idx], '.')
+        plt.xlabel('Lat')
+        plt.ylabel('Long')