Evidence of E-petition inaccuracies

Gentlemen, I can prove there is counting fraud when it comes to the revocation Article 50 E-petition. Attached is a screenshot of the claimed number of signatures. At time of writing, it claims 2,281,316 have signed, but this is actually impossible.

On the page, they supply a link to the raw JSON data used for calculating signature numbers, found here:

https://petition.parliament.uk/petitions/241584

On viewing the map, I had a sneaking suspicion the numbers didn't tally, so I converted the JSON data into CSV, and lo and behold, it doesn't.

CSV data for signature count by constituency:

https://pastebin.com/AxszvGjy

CSV data for signature count by country:

https://pastebin.com/piaG6Qaz

In the first CSV, the number of voters voting from constituencies sums up to 1,238,638

Which is evidently less than half of the number on the e-petitions page. But what about votes from other countries? If we factor in the countries (note: some country names had unicode characters and these had to be omitted), we find it totals to:

1,311,334. Aha, you might say, the other half! Except... 1,261,367 are counted as coming from the UK. Not only is there a DISCREPENCY between the first CSV and the second, but it also means the number of votes from countries is actually less. If we subtract the UK votes from the country list (already counted), we get a mere:

49,967votes

49,967+ 1,238,638 is not "2,281,316" signatures, dear readers... it is only 1,288,605.

Over half of those votes counted on the main page don't even exist!

Check the CSV data yourself.

Python 2.7 Code used to analyse the JSON data I wrote can be found here:

https://pastebin.com/09zBK677

It's supporting FileIO.py library can be found here:

https://pastebin.com/VKcBi6em

Verify it yourself people. It's evidently wrong.