First, Thanks all for the support and the blog mention.

I was out of home in the last weeks, and I will be busy for the next two weeks also. So, I had a little progress and few opportunities to answer the comments (the bad and good ones).

The hardware still running, but, since I'm out of home most of time, I did few measuments. Because of that, I think it will be important to make the portable hardware soon as possible. However, the Lightblue Blue still at customs.

Within the small data that I already have, I tried some regression algorithms, one of the best is the Self-organizing map. I will study it and try to use this algorithm in the next weeks to do some previsions and try to see if it works. As a self-leaning algorithm the previsions would become better as the time advances. Until now I used orange to do the data visualization and regressions.

I plan to do some changes in the next hardware versions. It will be portable and as I'm not totally happy with the noise level, I will try to improve it. First, I will try to put the photodiode and the ADC in the same board without cables, and with a metalic shield.

I also brought some laser diodes. While the LED has emissions with a width of 150nm, the laser diode has only 5-10nm. I brought 650, 808 and 1610nm laser diodes. With a narrow emission, I think it will be more selective to the glucose (and 1610nm is better than 1550nm in theory)