WARNING – liveblogging. Prone to error, inaccuracy and howling affronts to grammar and syntax. Posts will be improved over the next 48 hours
Google Docs notes for this session
(Alistair Rae introduced the ideas for this session in a blog post: How to map everything (but you definitely shouldn’t)
The age of open data has created an George Mallory approach to geo data: why did you create it? Because it’s there. Alistair has created loads of maps just because he could. But “why?” needs to be asked more.
Do we need a geodata page for data journalists? – with common boundaries, generalised, she, lml, geojson and other file formats.
There are lack of good open source geodata file formats. Shapefile documents are handy because they’re packaged – but they’re larger abandoned as a developing format. Do we need to be moving forwards? if geojson isn’t enough, is there a way to adapt it and move it forwards. Do people want services rather than data? One attendee converts everything into spatial databases. Flood alerts are published in geojson format – and there’s been loads of use, and no negative feedback. Practical experience suggests that it’s perfectly acceptable – as long as the information doesn’t get too big.
Json is used in RESTful APIs – so most people are willing to handle Json. QGIS is forcing us towards Shapefiles, still, as you need it for some transformations. QGIS saves geojson in a compact format without diffs.
There is at least one major council in the UK adding data to Open Street Map on a specific provider – because it’s cheaper than working with their outsourced provider.
Medical catchment data
Prescribing data is available as open data. That can be mapped to surgery postcodes – but not to catchment data, or police data. Can this be brought together? That catchment area data is a big problem. Some people have estimated them using a routing algorithm. It can be done – but the catchment areas just don’t exist. Somewhere there’s a GP via postcode map, because some services offer it. But could it just be done by geographic proximity. GPs are resistant to data release, because of the risk of it being used to assess performance.
Interesting that several people want a variety of boundaries/catchment areas: Police, GPs etc #odcamp
— SK53 (@SK53onOSM) May 14, 2016
Finding open data
Where do people look for open data? A pervasive problem. Term definition is one. Even the most well-describe data sets need time playing with them to understand them. Computer readable descriptions would be useful, rather than just paragraphs of text – but i would need to be standardise. There is a format in use on data.gov.uk.
As a community we could develop and crowdsource the information through approaches like Stack Overflow discussions (there is an Open Data StackExchange. It’s something we should tackle. We may be looking at it wrong – we index the data, not the problems. Connect people together who are, or have, worked on a particular problem.
Questions about #opendata? There is an Open Data @StackExchange Q&A website: https://t.co/sGkNqqzVgz#odcamp
— Giuseppe Sollazzo (@puntofisso) May 14, 2016
Data.gov.uk was amazing – five years ago. Now there’s a discussion happening about what it needs to be now, because it’s shit by today’s starboards.
Metadata is still seen as part of the publishing process, not part of the data management process. How do we sell that to data providers?
Ordnance Survey polygons sometimes shift if you try to use them at street level – they’re not designed for that level of accuracy. There are vast numbers of organisations that are totally geared up to using this system for absolutely everything – they only hit problems when someone gives them a KML file.
We need more simple “how to” guides. Some have be done – but aren’t ranking well in Google. Where can we put them to help people find them. Do we need some form of library that pulls this together?
Open Street Data obsessions
How detailed should Open Street Map data get? Should they add the trees? It depends what you’re interested in. Some people are keen on mapping trees, for example, because that can help with all sorts of research. Post boxes are really useful. Post boxes get moved – and the Post Office’s data doesn’t catch up quickly. They’re near commercial centres – so they indicate places that are interesting. Plus it lets people who spend all their time in front of screen get out at the weekend…
"If I wanted to find insects, I would map trees". Level of details of data collection is clearly use-dependent.#odcamp #opendata
— Giuseppe Sollazzo (@puntofisso) May 14, 2016
The only good map of cycle routes on Open Street Map – Sustrans doesn’t have it. West Midlands refused an FOI request on them for security reasons. But volunteer firefighters ore doing that mapping for very good reasons.