Recently, I was doing a project in computer science (involving modes of computation) that had a very specific time limit. Unfortunately, due to software problems, I was forced to obtain my data by hand. I went to substantial lengths to ensure the correctness of this data, but am worried that my credibility and the credibility of the data will be significantly decreased. What should I do to maintain credibility?
There is probably little to do but be honest about how the data was obtained. If it holds up under review then you have no issues. The past is the past.
However, if you have a chance you can try to obtain a better data set under more controlled conditions. This might be the core of a follow-up publication, either validating or invalidating the earlier result.
You learn something from just about every study. But if you are honest, it is hard to complain.