The main content of a NewsCred article.

Since the NewsCred data source delivers full-text news articles, it is useful to know the length of text we expect to recieve. The length is determined by NewsCred themselves and DataSift does not perform any truncation. A typical article is 2,000 to 3,000 characters long but we have seen cases where the length exceeds 40,000 characters.

Due to the length of NewsCred full-text articles, you can write some interesting DataSift filters by combining the contains_near operator with the newscred.articles.text target. Here, the full power of the operator emerges. Tweets, for example, are so compact that no word in the content can be very far from any of the others. However, with NewsCred we can create some more powerful analysis using nothing but contains_near.


  1. Filter for news articles that mention Disney theme parks:

    newscred.article.content contains_any "Disneyland, Disney World, Tokyo Disney,
                          Disneyland Paris, Hong Kong Disneyland, Shanghai Disney"

  2. Use DataSift's contains_near operator to filter for articles that mention "Cancun" and "vacation" within 200 words of each other:

    newscred.article.content contains_near "vacation, cancun:200"

  3. A more intricate DataSift stream with contains_near, filtering for three pairs of keywords that must appear close to each other but with no requirement that they should be close to any of the other pairs:

    newscred.article.content contains_near "Yahoo, Microsoft:400" and
    newscred.article.content contains_near "staff, redundancies:200" and
    newscred.article.content contains_near "email, advertising:200"

Resource information

Target service: NewsCred

Type: string

Array: No

Always exists: No