in 2009 each computer mined a block(solo mining, one cpu=one vote)
you can look at the first 1000 blocks (jan 9- jan19) so average 100 blocks a day
you can look at the next 2016 blocks (upto 3016) (jan19- feb4) is 17 days so average 118 blocks a day
and so on
and so on
then you can look at the "patoshi extranonce pattern" data, or just get extranonce from your own blockchain data
to see which blocks got solved in sequences (meaning same computer, progressing one block to next)
and you can see how many computers were involved over time via seeing how many different sequences were operating
for instance in the first 79 blocks, block 12 and block 65 was out of sequence, suggesting 1 computer mined 77 of the first 79 blocks which was 2 days so 39 blocks a day max mining in first couple days
you can then extrapolate more data for other days after it using the patoshi data/extranonce to kind of guess how many different systems were operating in 2009
for instance
i just thought of random block 5000.. and looked at the what day it appeared(feb20th)
and found first block of that day 4932 and last block of day 5048 (uk time)
meaning 116blocks that one day
there were 5 separate sequences happening that day(5 computers)
where the best sequence solved 89 of the 116 blocks
so that random day i picked, had one computer getting 89*50=4450 coin
heres a screen shot of part of that days sequences of extranonce(red is blocknumber) to show 5 columns(computers) of different sequences
