Thousands by Thousands
November 26, 2010
Yang recently re-jigged the MODIS and MISR estimates for the GBD study. This is a good thing for the project — the estimates are likely more realistic, and Yang was able to (1) back-estimate values for 1997-2001 and (2) fill other missing values in the data. This is a slight pain for me, because I have to go back and re-do quite a lot of work with the new values and across the new timeline. No big deal, but somewhat time-consuming since I haven’t given much thought to this code over the past 4 months.
Over the past two weeks I’ve been re-running everything and today I finally output a final result — totally weird when compared to our previous results, which it shouldn’t be. Hmm. A little investigating led me to realize that Yang had also re-jigged his factor of 10. Where the old MODIS and MISR values needed to be multipled by 1000, the new ones did not. I didn’t look carefully at the data at that stage, and by the time the fused estimate had been calculated the error wasn’t glaring, since it had been averaged into a bunch of other values. Again, no big deal — just another day in data analysis.
