Publication Overview 2017

published 2018-01-19

The whole world of research seems to be obsessed by scientometry. Unfortunately, questions like "how many papers" or "what is the impact factor" replace questions, which may bring truly new advances and insights. The discussion with my colleagues are full of "papers" and "citations". Having a fresh experience from Germany, I feel the contrast even stronger here in Czechia. No wonder, the reason is clear: in the Czech Republic, the science funding depends strongly on the scientometric criteria of the applicant. Certain adaptation to this external pressure is inevitable for any scientist who really means it, now matter one agrees or disagrees with it. Here is my piece of the obsession. During 2017, I monitored my Google Scholar account using a simple python script. I provide a few observations, which attracted my attention.

The upper plot (black line) shows the overall number of citations of my papers. The increase is pretty much linear with a jump/step around the 25th week. I have two hypotheses about the jump. Either there was a change in indexing new publications by Google or other publishers (like a university or an institute), or there was an increase in number of publications due to summer vacations. I haven't found any news about Google Scholar indexing changes from that time and was too lazy to check the actual time evolution of the number of publications in the related fields. In any case, the slope changed after the jump from roughly 4 to roughly 3. Hence, the estimate for the end of 2018 is around 800 citations.

The lower plot (color lines) shows the citation for each of my papers separately. Three papers appeared during 2017 and they have not collected many citations so far. Two of them are (semi)reviews (J Med Chem and Current Opinion), which promises a faster and more pronounced response. The most cited paper from the beginning 2017 (JCTC on halogen force fields) dropped down by one place in the end of 2017 being beaten by the Chem Rev on halogen bonding. On average, this review was cited more than once every week overcoming 100 citations in December.

It is perhaps not crystal clear from the plot, but some papers lost a few citations from time to time. I attribute this effect to Google Scholar way of dealing with the papers. Google Scholar often contains publication references in several variants at the same time (e.g. arXiv preprint + original publication). Apparently, there are correcting mechanisms which cluster equivalent items into groups. This may consequently decrease the citation impact.

The python script is still up and running and I may or may not bring another annual report next year. This is where my obsession ends. It was Wilfred van Gunsteren who said that we should not count the papers. We should rather read them.