Mashing up
Michael Brunton-Spall
michael.brunton-spall@guardian.co.uk
@mibgames

About Us
Online since 1995
250M+ pages per month
30M+ visitors per month

1995 - Guardian Online

1999 - Guardian Unlimited

1999 - Removed the registration wall

2007 - Rebuild and Redesign

2007 - Rebuild and Redesign

2007 - Rebuild and Redesign

Developing The Guardian

The Hackable Guardian
Url Hacking
Keyword Combiners
RSS Feeds

Url Hacking
http://www.guardian.co.uk/
[section]/[keyword]
technology/internet
[section]/all
environment/all
[publication]/[date]/[newspapersection]
theguardian/2009/jun/11/technologyguardian
profile/[name]
profile/bobbiejohnson

RSS Feeds
RSS Everywhere!
[section]/[keyword]/rss
technology/internet/rss
[section]/all /rss
environment/all/rss
[publication]/[date]/[newspapersection]/rss
theguardian/2009/jun/11/technologyguardian/rss
profile/[name]/rss
profile/bobbiejohnson/rss

Full Fat!
http://www.flickr.com/photos/snapperwolf/

Url Combiners
A Logical AND
Almost any combination from before
technology/internet+profile/bobbiejohnson
theguardian/technologyguardian+technology/internet
Except things that aren't actual tags:
Dated newspaper sections
theguardian/2009/jun/11... + anything = 404
the all page
.../all + anything = 404

Problems with this approach
Fragile
No Discovery
Not well documented
Copyright Issues

Build applications with the Guardian

Open Platform
Opening up how we work with people both internally
and externally
A suite of services enabling partners to build
applications with the Guardian
Content API
Data Store

Data Store
A directory of useful data curated by Guardian editors

Data Store
A directory of useful data curated by Guardian editors

Data Store

Data Store

Content API
A service for selecting and collecting content from the
Guardian for re-use

Content API
A service for selecting and collecting content from the
Guardian for re-use

Content API
Search content

Content API
Xml and Json

Content API
Tag Metadata

Content API
Full Article Body

Content API
Filters

So why?
guardian.co.uk has an amazing amount of
quality content
Incredible amounts of meta-data, curated by
editors
Aim to allow the guardian to become a rich
source of facts and journalism for the web

How?
Backed of our search platform
Provides access to 10 years of article content
and metadata
Supports multiple output formats: XML, JSON,
ATOM
Supported free text search across content
Search for keywords
Guardian supported api projects in Java, PHP,
Python and Ruby
Community supported api projects in Perl,
ActionScript and Coldfusion (and probably
more?)

Pricing

Pricing
FREE!

How free is free?
You can publish full articles from the guardian on your
website
5k queries per day limit
24 hour maximum cache lifetime
Online support
Partner with us on advertising

Beta trial
Limited number of keys
Collecting feedback
Will open more widely at end of beta program

How do I use it?
Home page
http://www.guardian.co.uk/open-platform
Sign up for an API key at
http://guardian.mashery.com
Use the API Explorer at
http://api.guardianapis.com/docs/
Use the python library at
http://code.google.com/p/openplatform-python/

What can I do with it?
Search our tag hierarchy
Find content by tag
Find content by search terms
Display on your website!

MP's Expenses

MP's Expenses
Written in Django
Hosted on EC2
Easy to modify

MP's Expenses
An MP's Page

View
def mp(request, id):
mp = get_object_or_404(MP, pk = id)
...
return render(request, 'mp.html', {
'mp': mp,
'documents': mp.documents.all(),
'top_users': top_users[:5],
})

View
def mp(request, id):
mp = get_object_or_404(MP, pk = id)
...
return render(request, 'mp.html', {
'mp': mp,
'documents': mp.documents.all(),
'top_users': top_users[:5],
})

View
def mp(request, id):
mp = get_object_or_404(MP, pk = id)
...
return render(request, 'mp.html', {
'mp': mp,
'documents': mp.documents.all(),
'top_users': top_users[:5],
})

Using the Guardian API
Getting all articles about an MP
results = client.search(q='"%S"' % (name))
returns a paginating iterator
for x in results:
print x['headline']
Shows only first 10 by default

Get the client
def mp(request, id):
mp = get_object_or_404(MP, pk = id)
...
client = Client(settings.GUARDIAN_APIKEY)
return render(request, 'mp.html', {
'mp': mp,
'documents': mp.documents.all(),
'top_users': top_users[:5],
})

Make the request
def mp(request, id):
mp = get_object_or_404(MP, pk = id)
...
client = Client(settings.GUARDIAN_APIKEY)
return render(request, 'mp.html', {
'mp': mp,
'documents': mp.documents.all(),
'top_users': top_users[:5],
'articles': client.search(q='"%s"'%(mp.name)),
})

Update the template
<div id="about-mp">
<ul>
{% for article in articles %}
<li>
<a href="{{article.webUrl}}">{{article.headline}}</a>
</li>
{% endfor %}
</ul>
</div>

And the result!

Additions?
Use Memcached to cache the response from guardian API
Don't create a new client each time
Article bodies

Who else is using it?

Who else is using it?

Who else is using it?

Who else is using it?

Public API Key?
jbynv3fwdp8ju5625mt2axw3
Only 5k queries a day
Will be closed on Monday 6th July
Will be closed if abused
Strongly encourage you to sign up for your own key

Questions?