Frequently asked Questions
How are petition classifications generated?
The current e-petition website only classifies petitions according to two criteria:
- Open or rejected
- Submission/closing dates
For the purposes of this website, I wanted to extend that to make petitions searchable and browsable by more categories. The additions I have made are:
- Split "open" into "open" and "closed" (depending on whether the closing date has passed)
- Autmatically taggged petitions with common keywords
- Automatically extracted links to legislation referenced in petitions
- Created four primary categories for browsing petitions:
- "Top petitions" - those with 1,000 signatures
- "Big hitters" - open petitions with the most signatures in the past week
- "Recent arrivals" - newly approved petitions
- "Tumbleweed corner" - a more left-field look at the petitions which failed to garner any support at all!
- Added signature count rank and "diversity index" values for all petitions
What happened to the signatures by postcode?
The previous version of the government e-petitions website published stats by postcode district for petitions with over 1,000 signatures. The new version replaces that with a count per constituency for all petitions. As the postcode states are now only of historic interest, they are no longer shown on the site.
What is the "Diversity index"
The diversity index attempts to give a measurement of how geographically dispersed signatories are to a petition. The higher the number the more widely spread signatories are, while lower numbers mean that signatories are concentrated in a smaller number of locations.
The index is calculated as a Gini coefficient, which is a widely used measurement of inequality. For the purposes of this site, the base values are reversed and then multiplied by 100 to give a percentage diversity score. (On a normal Gini coefficient, low numbers mean greater equality, but I felt that making high numbers mean more equal distribution made more sense in the context of this website).
A low diversity index is not necessarily a bad thing; many petitions relate specifically to local isses and it is only to be expected that they will have most of their signatories from the affected area. But where petitions are about a national issue, the diversity index can be a good indicator as to whether it has broad-based appeal or not.
How are petition tags generated?
Tags are automatically extracted by software, looking for commonly recurring keywords. Words are stemmed (using the Porter Stemming Algorithm) before being matched, so related words will match the same tag. For example, "digital", "digitally" and "digitise" will all be tagged with "Digital".
Because the tagging is automated, it may generate tags that are not directly related to the subject of the petition. The tagging system is being refined in the light of experience, so it isn't necessarily a fixed list.
How are links to legislation generated?
As with tags, relevant legislation is automatically extracted by software. The difference is that there is a fixed list of Acts of Parliament and Bills before parliament, so any match will always be reliable. But it may miss those which are misspelled or referred to by an abbreviation rather the full title - the system attempts to spot some common typos and abbreviations, but not all of them.
The legislation matching system doesn't attempt to distinguish between multiple Acts from different years with the same name, even where the petition's author has specified the year.
What gives you the right to use my petition on your site?
The content of the government e-petitions website is © Crown Copyright. The terms and conditions of the site permit re-use on other websites, providing the source and copyright are acknowledged. This website complies with those conditions.
Why is there advertising on this site?
Because this site isn't funded by the government, or an educational establishment, or a charity, or a large corporation, and I still have hosting costs to pay.