Data are contained in millions of spreadsheets, and everybody working with Excel-files has already experienced some peculiarities. This is not only a problem for the exchange of data, since often the formulas used for calculating the values in fields are exported and not the values themselves. Snafus and bugs also an obstacle for sciences like economics or genomics, because the complex values at times are misinterpreted; “SEPT2”, which corresponds to the gene Septin 2, is read by Excel as “September 2nd”. A recent study in the journal Genome Biology investigated papers published between 2005 and 2015, and identified an astonishing proportion of spreadsheet-related errors in them. This poses a real challenge for subsequent scientists who try to build on previously published research results.
Jörg Lehmann, FUB
Results of statistical operations are far from being intuitive. One needs a lot of experience to reliably judge computed results, and most people would listen to the advice given by experts in the field. This makes a story presented by the BBC all the more remarkable; it is about a student who took over the task to check the calculations of two Harvard professors. He was not only unable to replicate the results that Carmen Reinhart and Ken Rogoff published; it was much more: he spotted a basic error in the spreadsheet and was thus able to correct their research paper called “Growth in a Time of Debt”. That may sound like a modern myth; but sometimes politicians base their decisions on papers like these.
Jörg Lehmann, FUB
The KPLEX team was pleased to present their project at the Big Data PPP Info Day in Luxembourg, 17-18 January 2017. Although categorised by our panel chair as ‘different,’ the project concept received a good response and a number of good questions and leads for future work. We look forward to contributing more in the future to the widening of perspectives and development of RRI in big data research.
This article in the Guardian summed up some of the thinking that went in to the KPLEX research programme. What good is all this data if we no longer have faith in it?
The KPLEX partners started off 2017 with a toast to the official launch of our project, which commenced on 1 January 2017. We look forward to spending the next 15 months applying humanities knowledge and contexts to the investigation of what we reveal (and what we hide) when we speak about ‘big data.’