Download Taxi data if not yet cached

367eb404 · Jon Mease · 95bb8ead · 367eb404
Commit 367eb404 authored 7 years ago by Jon Mease
--- a/notebooks/nyc_taxi_selection.ipynb
+++ b/notebooks/nyc_taxi_selection.ipynb
@@ -14,6 +14,7 @@
   "metadata": {},
   "outputs": [],
   "source": [
+    "import os\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import plotly.graph_objs as go\n",
@@ -26,7 +27,24 @@
   "metadata": {},
   "outputs": [],
   "source": [
-    "df = pd.read_pickle('data/nyc_taxi.pkl')\n",
+    "pkl_path = 'data/nyc_taxi.pkl'\n",
+    "if os.path.exists(pkl_path):\n",
+    "    print('Loading saved dataset file... ', end='')\n",
+    "    df = pd.read_pickle(pkl_path)\n",
+    "    print('done')\n",
+    "else:\n",
+    "    print('Downloading and saving dataset (thanks to the datashader project for making this example dataset available!)... ', end='')\n",
+    "    df = pd.read_csv('http://s3.amazonaws.com/datashader-data/nyc_taxi.zip', compression='zip')\n",
+    "    df.to_pickle(pkl_path)\n",
+    "    print('done')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
    "df.head()"
   ]
  },
@@ -244,7 +262,7 @@
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
-   "version": "3.6.6"
+   "version": "3.6.5"
  }
 },
 "nbformat": 4,

 %% Cell type:markdown id: tags:
 ### Load NYC Taxi data
 We're going to look at the distribution of trip distances for a potion of the NYC taxi data set for January 2015.
 %% Cell type:code id: tags:
 ``` python
+import os
 import pandas as pd
 import numpy as np
 import plotly.graph_objs as go
 from ipywidgets import VBox, HBox
 ```
 %% Cell type:code id: tags:
 ``` python
-df = pd.read_pickle('data/nyc_taxi.pkl')
+pkl_path = 'data/nyc_taxi.pkl'
+if os.path.exists(pkl_path):
+    print('Loading saved dataset file... ', end='')
+    df = pd.read_pickle(pkl_path)
+    print('done')
+else:
+    print('Downloading and saving dataset (thanks to the datashader project for making this example dataset available!)... ', end='')
+    df = pd.read_csv('http://s3.amazonaws.com/datashader-data/nyc_taxi.zip', compression='zip')
+    df.to_pickle(pkl_path)
+    print('done')
+```
+%% Cell type:code id: tags:
+``` python
 df.head()
 ```
 %% Cell type:markdown id: tags:
 Extract only those trips that carried at least 4 passengers. Also discard outlying trips fo more than 10 miles for visualization purposes.
 %% Cell type:code id: tags:
 ``` python
 df_cleaned = df.loc[np.logical_and(
    df.passenger_count >= 4,
    df.trip_distance.between(0, 10, inclusive=False)
 )]
 len(df_cleaned)
 ```
 %% Cell type:markdown id: tags:
 We're left with almost 1.2 million data points
 %% Cell type:markdown id: tags:
 ## Distribution of trip distance
 %% Cell type:markdown id: tags:
 Initialize an empty figure with fixed x-axis range
 %% Cell type:code id: tags:
 ``` python
 fig1 = go.FigureWidget(layout={
    'xaxis': {'range': [-0.1, 10]}
 })
 fig1
 ```
 %% Cell type:markdown id: tags:
 Add a histogram trace with predefined bins (This take a couple of seconds)
 %% Cell type:code id: tags:
 ``` python
 hist = fig1.add_histogram(x=df_cleaned['trip_distance'],
                          xbins={'start': -0.05, 'size': 0.1, 'end': 10})
 ```
 %% Cell type:markdown id: tags:
 Update the axis titles (This happens immediately)
 %% Cell type:code id: tags:
 ``` python
 fig1.layout.xaxis.title = 'Trip distance (mi.)'
 ```
 %% Cell type:code id: tags:
 ``` python
 fig1.layout.yaxis.title = 'Frequency'
 ```
 %% Cell type:markdown id: tags:
 ### Plot pickup locations
 %% Cell type:markdown id: tags:
 Create an empty figure with hidden axes
 %% Cell type:code id: tags:
 ``` python
 fig2 = go.FigureWidget(
    layout={'width': 400, 'height': 400, 'hovermode': False,
            'xaxis': {'tickvals': []},
            'yaxis': {'tickvals': []},
            'margin': {'b': 0, 't': 0, 'l': 0, 'r': 0}
           })
 fig2
 ```
 %% Cell type:markdown id: tags:
 Add a `scattergl` trace of the `x` and `y` coordinates of all 1.2 million pickup locations
 %% Cell type:code id: tags:
 ``` python
 scatter = fig2.add_scattergl(
    x=df_cleaned['pickup_x'],
    y=df_cleaned['pickup_y'],
    mode='markers',
    marker={'size': 4, 'opacity': 0.1})
 ```
 %% Cell type:markdown id: tags:
 Constrain the aspect ratio so that view isn't distorted on zoom
 %% Cell type:code id: tags:
 ``` python
 fig2.layout.yaxis.scaleanchor = 'x'
 ```
 %% Cell type:markdown id: tags:
 ## Install Selection Callback
 %% Cell type:markdown id: tags:
 Install selection callback function to update the trip distance histogram based on only the selected trips
 %% Cell type:code id: tags:
 ``` python
 def update_hist(trace, points, state):
    if points.point_inds:
        hist.x = df_cleaned['trip_distance'].iloc[points.point_inds]
    else:
        hist.x = df_cleaned['trip_distance']
 scatter.on_selection(update_hist)
 ```
 %% Cell type:code id: tags:
 ``` python
 ```