Understanding Public Opinion 

Meme-machines has devised a better way to find out what people are talking about. It bakes privacy into the solution , introduces no bias, and is competely based on public data.

Corporations and political parties struggle to meet this need using existing techniques because they start from pre-set questions, which frame the debate, and can therefore miss underlying concerns.

The solution being marketed by Meme-machines offers analysis of open-ended questions. This enables organisations to ensure that public concerns are identified, and that none pass under the radar. As an example, the UK Brexit debate is presented here.

Stay tooned for US midterm coverage, coming soon!

Question: What do they Tweet about Brexit, and where?

Expand these sections to see how the map was made.


The Parliamentary geographical boundaries, used to calculate search circle location and radius. The more opaque the circle appears, the more it contains popular themes. ( Source )

The Brexit vote calculated by Parliamentary constituency, used to select and size icons ( Source )


For each constituency a search is performed on Twitter using a geographical filter, which limits results to those coming from within a circular area.

For example here is the search request for Aberavon

geocode: 51.5884895, -3.7047501, 7: km AND brexit

Mine and store

Full details of the proprietary mining process, along with academic papers and references, are are available here or you can use your Twitter account to explore Sherlock for yourself, use the login button at the top of this page

In brief, the search results have all common words removed, and only those that combine with others are retained as results.

The search results are stored in Redis

Along with Twitter features




      Sherlock finds the common interests that bind groups together, and needs no further information than the words written. It doesn't need, use,  or store private data. Privacy is baked into the design.

      Always up to date

      If an earthquake occurs, or a new political sensation invades the press, new names will baffle supervised text-mining techniques. Sherlock works from a list of what to ignore, so new stuff gets handled automatically, just so long as it is frequent.


      Sherlock has been tried on more than thirty languages, including Arabic and Russian. It is completely grammar agnostic, and will work on any language, human or computer, for which a stop word list can be provided.


      Sherlock tries to be as autonomous as possible, requiring just two parameters, how robust the patterns need to be. and how many words will be enough.  For example, the survey takes 75% pattern confidence, and 20 words

      © Copyright 2014-2019 meme-machines.com - All rights reserved.