The public has spoken - exploring e-petitions

Petitions for Action - Verbs used in e-petitions

E-Petitions as an alternative to consultation

The e-petitions on Number10.gov are fascinating.  As we've come to expect from mySociety projects, "the e-petition system has been designed to be transparent and trustworthy".  The petitions provide an unbiased perspective on the view of the UK public and a valuable evidence base for policy makers.

Although the feedback is free from the constraints of normal consultation (where the questions may be  framed unintentionally) it is unstructured and this makes it difficult for policy makers to use.  Hope is not lost however, help is at hand in the form of semantic analysis.

Extract information from e-petition data with semantic analysis

The image above was constructed by annotating the text in the titles of the petitions.  We searched for verbs and extracted the pruned form of those verbs (e.g. "increasing" becomes "increase") along with the number of signatures to the petition where the verbs were found.  The words are weighted according to the total number of signatures.  The visualisation was prepared using IBM's Many Eyes (note that data submitted to Many Eyes can be viewed by anyone on the internet).

Annotation can also be used to find organisations and places.  The results of those analyses are presented in the lists below.

Beware of junk data

Although the openness of e-petitions is a virtue, it does mean that there is a lot of junk to sift through.  Only 1.3% of open petitions have enough signatures (at least 200) to qualify for consideration by Government.  Clearly the other petitions have some value even though they may be minority concerns.  Indeed the number of signatures is not a perfect measure of quality or relevance.

Using the evidence in policy development

• Keep track of petitions that are made regarding your organisation or place.  We were able to identify 352 references to organisations, 36 of which were Councils.  Get in touch for some tips on monitoring the web for references to your organisation.
• Identify issues that are relevant to your objectives.  We recently employed semantic analysis that helped Investor Developers at the East Midlands Development Agency forecast demand for support by extracting information from visit reports.  Petitions provide evidence of needs (or at least demands) and issues (learn from others mistakes).
• Gauge public opinions on existing policy.  Congestion-charging, War, Bank-bailouts - they're all subjects of petitions.  While many of the petitions express peoples' feelings on these issues there are others that contain concrete suggestions for policy improvements.