# Help finding a formula to fit the data - both axes are logarithmic?

I'm an electrical engineer trying to come up with a formula to turn a measurement of light level into lux. The photoresistor changes resistance logarithmically. Lux is also logarithmic. So the data I have collected looks like this:

My math is failing me right now; is there a formula for me to convert the raw readings into Lux that would fit this curve? How would I find it? The exponential regression for my two samples does not fit closely enough to be useful.

Raw Lux
0   0
68.5    30
81  40
88  55
91.5    70
95  85
98.5    100
100.5   115
102.5   130
103.5   140
104.5   155
106 170
109 320
110.5   460
113.5   600
114.5   750
115.5   900
116 1000


• Wouldn't this be better for Stats.SE? – Mark Fantini Aug 25 '15 at 17:21
• I don't know. When I was typing the title of this question there were a lot of similar ones. – Bob Baddeley Aug 25 '15 at 17:26