Odsc london data science bootcamp with pixie dust

@DTAIEB55
Data Science Bootcamp at ODSC Europe 2017
Analyze Data and Build a Dashboard with PixieDust
London, October 2017
David Taieb
Distinguished Engineer
Developer Advocacy
IBM Watson Data Platform

©2017 IBM Corporation
@DTAIEB55
• How to use Jupyter Notebooks to load, visualize and
analyze data
• IBM Data Science Experience (DSX)
• PixieDust open source Python Library
• How to build a dashboard using PixieApps
• U.K Open Data
• Mapbox GL
What you will learn in this Workshop

@DTAIEB55
• The tutorial can be followed from a local Jupyter Notebook
environment. However, the instructions and screenshots here
walk through the notebook in the DSX environment.
• A corresponding notebook is available here:
https://github.com/DTAIEB/demos/blob/master/tutorials/ODS
C%20London%20UK%2BTraffic%2Banalysis.ipynb
For best results, use the latest version of either Mozilla Firefox or
Google Chrome.
Info before we start

@DTAIEB55
DSX is an interactive, collaborative, cloud-based environment
where data scientists, developers, and others interested in data
science can use tools (e.g., RStudio, Jupyter Notebooks, Spark,
etc.) to collaborate, share, and gather insight from their data
DSX

@DTAIEB55
Sign Up
• DSX is powered by IBM Bluemix, therefore your DSX login is same as your IBM
Bluemix login. If you already have a Bluemix account or previously accessed DSX
you may proceed to the Sign In section. Otherwise, you first need to sign up for
an account.
• From your browser:
1. Go to the DSX site: http://datascience.ibm.com
2. Click on Sign Up
3. Enter your Email
4. Click Continue
5. Fill out the form to register for IBM Bluemix

@DTAIEB55
From your browser:
1. Go to the DSX site: http://datascience.ibm.com
2. Click on Sign In
3. Enter your IBMid or email
4. Click Continue
5. Enter your Password
6. Click Sign In
Sign In

@DTAIEB55
You will need to create a noteboook to experiment with the data and a project to house your notebook.
After signing into DSX:
1. On the upper right of the DSX site, click the + and choose Create project.
2. Enter a Name for your project
3. Select a Spark Service
4. Click Create
From within the new project, you will create your notebook:
1. Click add notebooks
2. Click the Blank tab in the Create Notebook form
3. Enter a Name for the notebook
4. Select Python 2 for the Language
5. Select 1.6 for the Spark version
6. Select the Spark Service
7. Click Create Notebook
Create a blank new notebook

@DTAIEB55
You can also import the complete notebook
1. Click add notebooks
2. Click the From URL tab in the Create
Notebook form
3. Enter a Name for the notebook
4. In the Notebook URL enter:
https://github.com/DTAIEB/demos/ra
w/master/tutorials/ODSC%20London
%20UK%2BTraffic%2Banalysis.ipynb
5. Click Create Notebook
Skip this page if you prefer to start from a blank notebook.

@DTAIEB55
• When you use a notebook in DSX, you can run a cell only by selecting
it, then going to the toolbar and clicking on the Run Cell (▸) button.
When a cell is running, an [*] is shown beside the cell. Once the cell
has finished the asterisks is replaced by a number.
• If you don’t see the Jupyter toolbar showing the Run Cell (▸) button
and other notebook controls, you are not in edit mode. Go to the
dark blue toolbar above the notebook and click the edit (pencil) icon.
Make sure your Notebook is in Edit Mode

@DTAIEB55
Edit Mode

@DTAIEB55
PixieDust
You can find detailed info on PixieDust here: http://ibm.biz/wdp-pixiedust
In this section, we’ll show how to use the PixieDust Open Source Python library within
a Notebook to:

@DTAIEB55
Make sure PixieDust is at the latest version
• DSX already comes with the PixieDust library installed, but it is always a good idea to
make sure you have the latest version
• In the first cell of the notebook enter: !pip install --upgrade pixiedust
• Click on the Run Cell (▸) button
• After the cell completes, if instructed to restart the kernel, from the notebook toolbar
menu:
• Go to > Kernel > Restart
• Click Restart in the confirmation dialog
Note: The status of the kernel briefly flashes near the upper right corner, alerting when it
is Not Connected, Restarting, Ready, etc.

@DTAIEB55
Import PixieDust
• Before, you can use the PixieDust library it must be imported into the notebook
• In the next cell enter:
import pixiedust
• Click on the Run Cell (▸) button
Note: Whenever the kernel is restarted, the import pixiedust cell must be run before continuing.
PixieDust has been updated and imported, you are now ready to play with your data!

@DTAIEB55
Import U.K traffic accidents data
It would be a good idea to spend a few minutes browsing the San Francisco open data
site: https://data.gov.uk
• Load the data: With PixieDust, you can easily load CSV data from a URL into a PySpark
DataFrame in the notebook.
• In a new cell, enter:
accidents =
pixiedust.sampleData("http://data.dft.gov.uk/road-accidents-
safety-data/dftRoadSafetyData_Accidents_2016.zip")

@DTAIEB55
Explore the Data
In the next few cells, we’ll use the PixieDust display() to interactively explore the data. This tutorial will
provide a few suggestions for which options to choose, but feel free to also explore on your own.
For example:
(Pie Chart - Options: Keys = PdDistrict, Values = IncidntNum, Aggregation = Count)

@DTAIEB55
Initial Exploration
Let’s start exploring our data by answering the following questions
• In which police district do the most traffic accidents occur?
(Pie Chart - Options: Keys = PdDistrict, Values = IncidntNum, Aggregation = Count)
• We can also dig one level deeper by clustering by how each accident was resolved:
(Cluster By: Resolution)
• On what day of the week do the most traffic accidents occur?
(Bar Chart - Options: Keys = DayOfWeek, Values = IncidntNum, Aggregation = Count)

@DTAIEB55
Results
Menu to select type of viz to display
Chart
Options
Dialog
Chart
Rendering area
Dynamic Options
(depend on the type
of chart and
renderer)
Renderer
Selector

@DTAIEB55
More data exploration and hypothesis
• Immediately, we can identify a couple of areas of interest in our data without
having to write a single line of code:
1. Most accidents happen in the Southern and Taraval police districts
2. Most accidents happen on Wednesdays and Thursdays.
• We can also see that our data needs some cleansing if we want to make
analysis easier:
1. The Time field needs to be converted into its time components
2. Rename the DayOfWeek values so they are rendered in alphabetical
order
• We should condense the outcome types of each traffic accident if we want to
see the most common resolutions of traffic accidents in each police district,
since the clustering above was unclear.

@DTAIEB55
Cleansing the data

@DTAIEB55
More exploration on the cleansed data
• Hypothesis: Do accidents in one police district result in more arrests than
other police districts?
(Bar Chart - Options: Keys = PdDistrict, Values = IncidntNum, Aggregation =
Count, Cluster By: Res)
• Question: How does the number of accidents change over the course of
the week?
(Line Chart - Options: Keys = DayOfWeek, Values = IncidntNum, Aggregation
= Count)

@DTAIEB55
Some more results

@DTAIEB55
What have we learned
• A few lines of code makes it a lot easier to see that:
1. Accidents in the Richmond police district are much more
likely to result in arrest than all other districts
2. The number of accidents peaks during the middle of the
week, but decreases afterwards as the week winds down.
Now let's focus on the Taraval police district using some friendly
SQL notation

@DTAIEB55
Let’s use Spark SQL to focus on TARAVAL
• In a new cell, enter the following code to create a temporary SQL Table so we
can then use regular a SQL query to select the accidents from TARAVAL.
• The results are then stored in a new Spark DataFrame
accidents.registerTempTable("accidents")
taraval = sqlContext.sql("SELECT * FROM accidents WHERE
PdDistrict='TARAVAL'")

@DTAIEB55
Exploration on TARAVAL accidents
• Question: Where in Taraval do most accidents happen?
(Map - Options: Keys = [X,Y], Values = IncidntNum, Aggregation = Count,
Renderer: mapbox, kind: chloropleth-cluster)
• Question: What time of day do most accidents occur?
(Line Chart - Options: Keys = Hour, Values = IncidntNum, Aggregation = Count)
Note: The Notebook comes with a public Mapbox access token. You can also create
your own free Mapbox access token at the following url:
https://www.mapbox.com/help/create-api-access-token/

@DTAIEB55
Map Visualization

@DTAIEB55
Let’s recap
Most of the results from looking at the accident times are unsurprising:
• Less accidents during very early morning (people probably sleeping)
• Steady increase in number of accidents during morning commuting hours
• Less accidents during mid-evening (people probably eating dinner)
• (Sadly) more accidents late at night.
The interesting thing here is the sudden spike in accidents during mid-afternoon (2-
3PM) - twice as many accidents happen during this two-hour window!
In the next section, we’ll bring more data to find more insights on the data. But
instead of writing more code in linear cells, we’ll build a dashboard using another
PixieDust feature called PixieApps

@DTAIEB55
What are PixieApps
• PixieApps are Python classes that let you write UI for your analytics
• Easy to build: mostly HTML and CSS with some custom attributes
• Leverage PixieDust Display visualization for charting
• With PixieApps you can:
• Create different html views with routes to invoke them
• Invoke Python Scripts from user interactions
• Run in the notebook cell output or in a Dialog
• and much more…
• Use cases:
• Dashboards
• Data Browsers
• Data Pipeline Management
Note: Please take the time to browse the PixieApp documentation page here:
https://ibm-cds-labs.github.io/pixiedust/pixieapps.html

@DTAIEB55
Import app package to start things off
Simple annotation to tell PixieDust it’s an app
Define the default route (no args).
Method will return the view’s html fragment
Html fragment for the view.
Allows Jinja2 template macros
Define a new route that triggers when
option clicked is set to true
set option clicked to true when button is
pushed, so that correct route is loaded next
Import app package to start things off
PixieApp Hello Word

@DTAIEB55
Placeholder div for displaying data
Display the output in the specified target
Entity binding: use the df passed by user
Allows binding of any entity created by the app
Specify Display options for visualization
Pass data to the app
PixieApp HelloWorld with Data

@DTAIEB55
Back to our U.K Data
• In analyzing the geographical data, we can see a couple of clusters where
accidents occur more frequently in Taraval - the southeastern corner
looks particularly crowded.
• Some useful questions to ask at this point are:
• Does crime has an effect on the number of accidents?
• Are there more accidents in these areas because more people speed
there
• Do traffic calming devices reduce the number of accidents?
Note: In the next section, we’ll download datasets from the San Francisco
Open Data Web site: https://data.gov.uk

@DTAIEB55
PixieApp Dashboard
Step 1: Create the skeleton
from pixiedust.display.app import *
@PixieApp
class UKDashboard():
def mainScreen(self):
return """
<div class="well">
<center><span style="font-size:x-large">Analyzing U.K Public Safety data with
PixieDust</span></center>
<center><span style="font-size:large"><a href="https://data.gov.uk"
target="new">https://data.gov.uk</a></span></center>
</div>
<div class="row">
<div class="form-group col-sm-2" style="padding-right:10px;">
<div><strong>Layers</strong></div>
</div>
<div class="form-group col-sm-10">
<div id="map{{prefix}}”/>
</div>
</div>
“””

@DTAIEB55
PixieApp Dashboard
Step 2: Create the map of accidents
@PixieApp
class SFDashboard():
return """
<div class="well">
<center><span style="font-size:x-large">Analyzing U.K data with PixieDust</span></center>
<center><span style="font-size:large"><a href="https://data.gov.uk"
target="new">https://data.gov.uk</a></span></center>
</div>
<div class="row">
</div>
<div id="map{{prefix}}”>
<pd_options>
{
"handlerId": "mapView", ”mapboxtoken":"pk.eyJ1Ijoic…","timeseries": "false",
"keyFields": "Longitude,Latitude","aggregation": "COUNT","kind": "choropleth-cluster",
"valueFields": "Road_Surface_Conditions","basemap": "light-v9","rendererId": "mapbox"
}
</pd_options>
</div>
</div>
</div>
“””
• What have we added:
• pd_entity: tell PixieDust which dataset to work on
• pd_options: Contains the PixieDust options for the map (see next slide for more info)

@DTAIEB55
Generating the pd_options for the Map
The best way to generate the pd_options for a PixieDust visualization is to:
1. Call display() on a new cell
2. Graphically select the options for your chart
3. Select View/Cell Toobar/Edit metadata menu
4. Click on the “Edit Metadata” button and copy the pixiedust metadata

@DTAIEB55
Initialize the Layers
def setup(self):
self.setLayers([…])
Note: setup is a special method that will be called automatically when the PixieApp is initialized.

@DTAIEB55
PixieApp Dashboard
Step 3: Create the GeoJSON customer Layers
from pixiedust.apps.mapboxBase import MapboxBase
@PixieApp
class SFDashboard(MapboxBase):
def setup(self):
…
self.setLayers([
{
"name": "Police Forces Area",
"url": "http://geoportal1-ons.opendata.arcgis.com/….."
},
{
"name": "Traffic Signals",
"url": "http://data.cyc.opendata.arcgis.com/datasets/…",
"type": "symbol",
"layout": {
"icon-image": "police-15",
"icon-size": 1.5
}
},
…
])

@DTAIEB55
PixieApp Dashboard
Step 4: Create the checkboxes from the layers
from pixiedust.apps.mapboxBase import MapboxBase
@PixieApp
class UKDashboard(MapboxBase):
…
@route()
return """
<div class="well">
<center><span style="font-size:x-large">Analyzing U.K Traffic data with PixieDust</span></center>
<center><span style="font-size:large"><a href="https://data.gov.uk/" target="new">https://data.gov.uk</a></span></center>
</div>
<div class="row">
{% for layer in this.layers %}
<div class="rendererOpt checkbox checkbox-primary">
<input type="checkbox" pd_refresh="map{{prefix}}" pd_script="self.toggleLayer({{loop.index0}})">
<label>{{layer["name"]}}</label>
</div>
{%endfor%}
</div>
<div id="map{{prefix}}" pd_entity>
<pd_options>
{"handlerId": "mapView",
”mapboxtoken":"pk.eyJ1IjoicmFqcnNpbmdoIiwiYSI6ImNqM2s4ZDg4djAwcGYyd3BwaGxwaDV3bWoifQ.d5Rklkdu5MeGAnXu1GMNYw",
"timeseries": "false","keyFields": "Longitude,Latitude","aggregation": "COUNT","kind": "choropleth-cluster",
"valueFields": "Road_Surface_Conditions","basemap": "light-v9","rendererId": "mapbox","rowCount": "600"
}
</pd_options>
</div>
</div>
</div>
"""

@DTAIEB55
Run the PixieApp Dashboard
You can find the final code for
the PixieApp in the complete
notebook :
https://github.com/DTAIEB/de
mos/blob/master/tutorials/ODS
C%20London%20UK%2BTraffic
%2Banalysis.ipynb

@DTAIEB55
Final Thoughts
• In this tutorial:
• We showed how PixieDust help notebook users (data scientists or not) to load
and visualize data without having to write code
• We also showed how to use notebooks in new ways thanks to PixieApps by
writing dashboard
• Next Steps
• Improve the analytics by finding new insights on the U.K Traffic accidents
dataset, or perhaps start a new notebook with a new dataset from
https://data.gov.uk
• Improve the PixieApp dashboard by adding new layers. Just find the geojson
link from the opendata site by clicking on the explore button (warning: not all
datasets come with a geojson link). Then just add a new entry in the layers
array

Odsc london data science bootcamp with pixie dust

More Related Content

Similar to Odsc london data science bootcamp with pixie dust (20)

Recently uploaded (20)

Odsc london data science bootcamp with pixie dust