Hi, I shared the article from Genome Biology to some acquaintances and students, more as a cautionary tale than as a "never use excel".* As every software, Excel is a tool: it is a user's duty to know how to use it and for what purposes. Knowing that there are problems with specific data, there may be workarounds (in the old times, you put double quotes around the text, I don't know if it still works). Or, you may decide to move to something else: command line tools, openrefine, databases etc. I'm not expecting to see someone loading a 8 Gb file into Excel, like I'm not expecting them to write a 100 lines program to perform a computation on a small data table, when such function is already implemented in their spreadsheet. But until they try to load such a file on a spreadsheet, or get introduced to documentation on the available functions, it is hopeless to expect a different behaviour.
I'm going to use the new article (and the old one from 2004) as examples of good and bad practices in managing data: knowing how the data was produced, how to transform it into something manageable and what are the limits of the software used in such a procedure should be key concepts lingering in learner's minds. Best, Giuseppe * Full disclosure: the last time I used Excel was at the end of last century, and I usually advocate the usage of open source software ;-) 2016-08-27 0:02 GMT+02:00 Emily Jane McTavish <[email protected]>: > Great points. > > I have a question about alternatives to excel for data input. > > Following this paper I have seen a lot of 'never use Excel' tweets, but that > seems to be ignoring a key step in real world data analysis pipelines. If > data is not coming straight off a machine, such as in ecological surveys, > behavioral experiments, meta-analyses of gene names, etc., those data need > to be put into a tabular, machine readable, format (e.g. CSV) somehow. I > don't think anyone is recommending using a text editor to do that. > Libre office calc and google sheets have many of the same autoformat issues > as Excel. (although that may be fixed in new versions of libre office?) > > I think when people say 'don't use excel', they often mean 'for analysis', > or 'for statistics'. But this paper demonstrates it is problematic for even > simple data input. I know what to recommend as alternatives in the former > cases, but not for the latter. Am I missing good alternative options here? > > > Thanks, > Emily Jane > > -- > Emily Jane McTavish > Assistant Professor > School of Natural Sciences > University of California, Merced > 5200 N. Lake Rd, Merced CA 95343 > [email protected], [email protected] > > > > > > On 08/26/2016 02:42 PM, Steven Haddock wrote: >> >> I was going to post that article too, but I dug into it (read the paper), >> and it is really just conversion of gene names (like SEPT5) in supplementary >> files. That was reported long ago as affecting some quantifications, but I >> would call it analytical errors as we have seen in the past. A bit of a >> tempest in a teapot, perhaps. >> >> Ironic twist, the paper provides a supplementary file listing all the >> gene-name errors they found, posted as an Excel file. >> >> -Steve >> >>> On Aug 26, 2016, at 14:26 , Maxime Boissonneault >>> <[email protected]> wrote: >>> >>> Hi everyone, >>> Some interesting content to use about how to not do science correctly >>> with a computer.... >>> >>> https://www.washingtonpost.com/news/wonk/wp/2016/08/26/an-alarming-number-of-scientific-papers-contain-excel-errors/ >>> >>> >>> Maxime Boissonneault >>> _______________________________________________ >>> Discuss mailing list >>> [email protected] >>> http://lists.software-carpentry.org/listinfo/discuss >> >> _______________________________________________ >> Discuss mailing list >> [email protected] >> http://lists.software-carpentry.org/listinfo/discuss > > > _______________________________________________ > Discuss mailing list > [email protected] > http://lists.software-carpentry.org/listinfo/discuss _______________________________________________ Discuss mailing list [email protected] http://lists.software-carpentry.org/listinfo/discuss
