Version of Churnalism.com for Africa

1

1. What do you propose to do? [20 words]

Extend churnalism.com to work with African news sources, press
releases and Wikipedia and work well with mobile platforms.

2. Is anyone doing something like this now and how is your project
different? [30 words]

Google has a news search facility, but it gives no indication of
duplication from other sources as an indicator of the originality of
the news.

3. Describe the real world challenge that you are trying to solve for
African media [50 words]

The cost of downloading data to mobile devices in Africa is high. If
readers can be assisted in discriminating between real news and
churnalism they can save money and be informed by facts rather than
adverts or government propaganda.

4. How and why will your solution work? [100 words]

A set of scrapers will ingest news and press releases from the most
popular sources in Africa. Superfastmatch
https://github.com/mediastandardstrust/superfastmatch is a technology
that can find all identical strings of characters in massive corpora.
For any news article or press release there may be a matching set of
documents that contain exactly the same text. The longer the
duplicated text the more likely churnalism/plagiarism has occurred. It
is informative to have an indication of this in ranked form before or
during reading an article.

5. Who is working on it? [100 words]

The Media Standards Trust completed a UK version of churnalism.com
last year. Sunlight Labs are currently working on the US version of
the site which includes a browser extension to give interactive
results as the user reads a news article. Donovan Hide is working on
improving the Superfastmatch engine to deal with a variety of
workloads for different analytical scenarios.

6. What part of the project have you already built? [100 words]

Most of the technology is already implemented for desktop browsers.
Mobile usage doesn’t permit browser extensions and typing in URL’s is
tedious, so some form of javascript badge on the news source site or a
central aggregating portal per country might be required to drive the
news article text through the engine. A related news option is an
extra feature that could be added, much like that provided by
http://www.moreover.com/ that might provide future revenue.

7. How would you sustain the project after the funding expires? [50 words]

The main costs of sustaining the project are a high RAM server in
an appropriate geographic location and the time to maintain the
scrapers if presentational changes occur. Related news could be a
revenue stream. Further grants might be another option.

Requested amount from ANIC: $60,000
Expected amount of time required to complete project: 6 months
Total Project Cost: $60,000


Name: Donovan Hide/Martin Moore
Twitter: @newsmatters
Organization: Media Standards Trust (in co-operation with Sunlight Labs)
Country: UK/US