search

datopian / bad-data

Examples of bad data, especially from government.
https://datahub.io/@rufuspollock/bad-data
22 stars 10 forks source link

Illegible spectrum #6

Open petermr opened 10 years ago

petermr commented 10 years ago

I published this on my blog about 6 years ago. I think this was from a * **** \ Chemistry journal but hold fire till I check. suppdata This was, of course , digital data in the spectrometer (perhaps 2^16 points)

rufuspollock commented 10 years ago

@petermr so was the point here that the data was not provided at all - just the graph?

petermr commented 10 years ago

Rufus, [Sit down before you read my reply.]

Number of spectra output from machines in digital form : approx 5 million / year (maybe 1 billion dollars)

Number of spectra published as PDF/png/tiff : approx 5 million / year (maybe 1 billion dollars)

Number of spectra published as data: approx ZERO /year

Data loss approx 1 billion dollars or worse.

On Sat, Nov 23, 2013 at 7:02 PM, Rufus Pollock notifications@github.comwrote:

@petermr https://github.com/petermr so was the point here that the data was not provided at all - just the graph?

— Reply to this email directly or view it on GitHubhttps://github.com/okfn/bad-data/issues/6#issuecomment-29139011 .

Peter Murray-Rust Reader in Molecular Informatics Unilever Centre, Dep. Of Chemistry University of Cambridge CB2 1EW, UK +44-1223-763069

rufuspollock commented 10 years ago

@petermr got you but for this specific example could we explain succinctly how data could have been provided - e.g. make explicit what the difference is between that graph and what could have been?

petermr commented 10 years ago

I have spent 10 years of my life trying to get chemists to do something as simple as save the digital output from the spectrometer. It's driven me wild. So has Henry Rzepa. When when the data is fraudulent, and proved to be so in public and when the simple device of providing the data would have averted this they STILL continue to print it out on paper.

When Tony Hey was invited to speak to Am Chem Soc he came away shell-shocked that such primitive life forms still existed...

Is there a positive way forward? The chemistry in the Imperial repository is closed for 5 years minimum - in case someone SEES it.

Maybe if David Willetts can tell the chemists ... but then I'd be a class traitor...

P.

On Sun, Nov 24, 2013 at 5:51 PM, Rufus Pollock notifications@github.comwrote:

@petermr https://github.com/petermr got you but for this _specific_example could we explain succinctly how data could have been provided - e.g. make explicit what the difference is between that graph and what could have been?

— Reply to this email directly or view it on GitHubhttps://github.com/okfn/bad-data/issues/6#issuecomment-29161128 .

Peter Murray-Rust Reader in Molecular Informatics Unilever Centre, Dep. Of Chemistry University of Cambridge CB2 1EW, UK +44-1223-763069