When preparing the light curve viewer for the WISE MEP table (roughly equivalent to the LSST ForcedSource), it is important to be aware of the precise definitions of the magnitude and magnitude-error columns in the table.
w1mpro_ep: Single-exposure W1 profile-fit magnitude, or the magnitude of the 95% confidence brightness upper limit if the W1 flux measurement has SNR<2. This column is null if no useful flux measurement could be made on this frame, or if the W1 frame was not used in the Multiframe processing.
w1sigmpro_ep: Single-exposure W1 profile-fit photometric measurement uncertainty in mag units. This column is null if the W1 profile-fit magnitude is a 95% confidence upper limit or if the W1 frame was not use in the Multiframe processing.
(Similar considerations apply for w2*, w3*, and w4*.)
When displaying this data in a time-series plot or other x-y plot, it would be very helpful to the science user if the plot symbol allowed the user to understand whether the w1mpro_ep value is a measurement or upper limit.
This could be minimally accomplished by ensuring that error bars (based on w1sigmpro_ep) are always displayed when available – and with the display of an error bar suppressed when the value in that column is null. In this case the absence of an error bar would be a cue to the user that the plotted value is an upper limit. While this preserves the available information, it requires the user to be aware of the WISE convention - it's not a universal one.
Therefore, it would be significantly preferable if we established a convention of a plot symbol to be used for upper limits, so that they remain intelligible even when the display of error bars is turned off. I would suggest that appropriate choices for upper limit symbols would be either a triangle or a "T" symbol, with the point of the triangle / stem of the "T" facing in the direction of fainter magnitudes (i.e., larger actual values of the variable). This convention could then also be used in general x-y plots, not just in time series.
It may also be useful to consider whether in table displays a visual cue could be provided to the user that a value shown is an upper limit.
When the value displayed is not a magnitude, but a flux (e.g., w1flux_ep), the AllWISE processing returns the actual best-fit value in all cases, which can therefore be a negative number for sources that are near or below the detection threshold for an image. Thus, for fluxes, no upper-limit symbol is required. The display of error bars (e.g., based on w1sigflux_ep) is essential for these negative values to be intelligible to scientific users.