We explore simple semi-analytic fits to the bolometric light curve of Gaia17biu/SN 2017egm, the most nearby hydrogen-deficient superluminous supernova (SLSN I) yet discovered. SN 2017egm has a quasi-bolometric light curve that is uncharacteristic of other SLSN I by having a nearly linear rise to maximum and decline from peak, with a very sharp transition. Magnetar models have difficulty explaining the sharp peak and may tend to be too bright 20 d after maximum. Light curves powered only by radioactive decay of 56 Ni fail on similar grounds and because they demand greater nickel mass than ejecta mass. Simple models based on circumstellar interaction do have a sharp peak corresponding to the epoch when the forward shock breaks out of the optically-thick circumstellar medium or the reverse shock reaches the inside of the ejecta. We find that models based on circumstellar interaction with a constant-density shell provide an interesting fit to the bolometric light curve from 15 d before to 15 d after peak light of SN 2017egm and that both magnetar and radioactive decay models fail to fit the sharp peak. Future photometric observations should easily discriminate basic CSI models from basic magnetar models. The implications of a CSI model are briefly discussed.

Superluminous supernova (SLSN) lightcurves exhibit a superior diversity compared to their regular luminosity counterparts in terms of rise and decline timescales, peak luminosities and overall shapes. It remains unclear whether this striking variety arises due to a dominant power input mechanism involving many underlying parameters, or due to contributions by different progenitor channels. In this work, we propose that a systematic quantitative study of SLSN lightcurve timescales and shape properties, such as symmetry around peak luminosity, can be used to characterize these enthralling stellar explosions. We find that applying clustering analysis on the properties of model SLSN lightcurves, powered by either a magnetar spin-down or a supernova ejecta-circumstellar interaction mechanism, can yield a distinction between the two, especially in terms of lightcurve symmetry. We show that most events in the observed SLSN sample with well-constrained lightcurves and early detections strongly associate with clusters dominated by circumstellar interaction models. Magnetar spin-down models also show association at a lower-degree but have difficulty in reproducing fast-evolving and fully symmetric lightcurves. We believe this is due to the truncated nature of the circumstellar interaction shock energy input as compared to decreasing but continuous power input sources like magnetar spin-down and radioactive 56 Ni decay. Our study demonstrates the importance of clustering analysis in characterizing SLSNe based on high-cadence photometric observations that will be made available in the near future by surveys like LSST, ZTF and Pan-STARRS.