Companies House offers free accounts data

A lot of data here but how do you find within that mass of data the accounts of the particular company that will be of interest in any given case?  What am I missing?  It sounds as though it may be worth continuing to pay £1 to go direct to what you need.

If the tool is indeed free then it would be possible to build a simple tool that allows a business to scrape the issued data for information. That way a business could get information about submitted accounts from anyone they were interested in (suppliers? customers? competitors?) the day after filing, if they were filed electronically.

A relatively simple activity could be scheduled five times a week which would pull "today's" items, unzip them on a server somewhere and check for any information necessary.

I presume someone will set up some sort of service which permits them to index the files if Companies House permits this. Any indexing service would be very simple to program and woudl require a comparatively small amount of disc space. Even if the files averaged to 50MB/day compressed, then this would only be 20GB/year so would be within the realms of web hosting costing approx. £5-10/month.

However, if Companies House doesn't permit any indexing or third party implementation of their data this might not be of interest to anyone without a handy and bored computer programmer in their organisation.

it is big

The data set is indeed free. It is about 100Gb uncompressed and contains 1.26 million sets of accounts. It took our computers around 3 days to process and index.

It is also about 15% xbrl which needs to be transformed to html.

How much indexing did you do that it took three days to process 100GB of uncompressed text?

Lots !

However I can list and view the accounts of companies incorporated on a Tuesday with intangible assets of over £5000 in the HP2 postcode area.

I just need to decide why that would be useful.

Wow, that seems brilliant. Are there tools for managing that or just a languages native xml parser, regex and an sql server?

(Which is what I was thinking. Python plus sql.)

Getting bit geeky

Getting a bit geeky for an accountancy forum. Feel free to PM me.

@dave

However I can list and view the accounts of companies incorporated on a Tuesday with intangible assets of over £5000 in the HP2 postcode area.

I just need to decide why that would be useful. 

Because you can write to them all in a mail merge and ask if the accountant has claimed tax on R and D expenses or correctly treated the amortisation of goodwill ( from tax or FRS 102 viewpoint) and would a free 1 hr review be of any use as it may save you some tax or being prosecuted by FRC.

....and now you have 50 new clients.( assuming you have the time to pop in and chat with them)

Or you can sell similar info to me in GU postcode ;o)

Tom

Ps

It always amused me that Epson printers are based in the HP post code area! Gotta be gutted your rivals name is on all the post