So if I have the following data available:
Person A
May 1 10
May 2 20
May 3 20
Person B
May 1 15
Person C
May 2 10
May 3 12
May 5 18
Person D
May 1 10
May 2 14
May 3 18
May 4 30
May 5 5
May 8 20
May 9 18
Person E
May 2 8
May 3 9
May 5 5
Whereby May 1,2,3 etc. are the working days where aforementioned persons worked on. The numbers represent the hours spent.
I would like to create a forecast of how many Persons are needed with this productivity that is available to reach the target of let's say 800 in the coming two weeks. How would I go about this..
First of all, you have not enough data. The working hours differ strongly between the days, so you can't really conclude things like "person A normally works 20 hours a day". Then there is the issue with different dates: For some of them you have data on days where there is non for others. Did they not work in these days? The person where you have the most days, person D, had really strange working times. 30 hours a day should be double checked...
Given all these problems, you should call "bullshit!" if you really got this data provided. However, if this is just a number example and you have more/better data, then that's a different thing.
To solve your problem in the latter case, first compute the number of hours a person X works each day, so take the average working time W(X). Then, given the different working times of X on different days, try to compute estimates for events like "X will work 10W(X) or more in total in the next 10 days", "X will work 9W(X) or less in the next 10 days",...
Once you have these values, you can write down the number of workers of each type needed and also give estimates on how likely it is that, given your data, these workers will in fact achieve the 800 hours. You might come up with different solutions, e.g. 30 workers of type X might manage with high probability, whereas 20 of type Y should suffice in case that you are lucky and they have their productive period coming up - from these solutions you then take what fits you best.