here it is:
http://www.alcula.com/calculators/statistics/linear-regression/
here's the results:
the first picture shows how to enter the inputs.
you can verify that the inputs were placed in correctly.
you can go to the website and re-run if you spot a problem with them.
i checked them out and i think they're ok, but it never hurts to check again.
the second picture shows the results.
sample size is 8 which is the number of data pairs that were entered.
mean of x and y are shown.
these are, i believe, the center values of the data on the graph.
the intercept is the y-intercept.
that's the value of y when the value of x is 0.
that doesn't mean much here because x = 0 would be over 2000 years ago.
slope is the average growth in internet users per year.
the regression line equation is used to find any point on the line and used to project future numbers of internet users assuming that the data continues the way it was in the data set shown.
that's a huge assumption and, for that reason, there usually are cautionary message indicating that the curve is fully valid only for the data points used to generate it.
projections in the future always have risk and don't always follow the general direction of the data that was history.
if you're reasonably certain that they won't deviate from the established pattern then it's ok to project based on the data, but always with caveats.
i will use a rounded version of the formula to do the projections.
the error factor should be pretty small.
i'll round to 2 decimal places which is probably more than required.
the formula becomes:
y = 146.95 * x - 293550.1
in the year 2012, the projected internet use is:
146.95 * 2012 - 293550.1 = 2113.3
to find the year in which the internet usage becomes 2549.2, set y = 2549.2 and solve for x.
you get:
2549.2 = 146.95 * x - 293550.1
add 293550.1 to both sides of the equation and then divide both sides of the equation by 146.95 to get:
(2549.2 + 293550.1) / 146.95 = x
solve for x to get:
x = 2014.97.
that would be late 2014.
to find the correlation coefficient using this calculator, just click on correlation coefficient and it will pop up.
you will find the correlation coefficient is equal to (r) = 0.9979812681594
that's very high which indicates a very strong fit which means you can determine with a good degree of accuracy what the data points lie just by using the equation.
there's another factor called R^2 which is R * R = .99596...
that's also quite high which indicates that most of the projection is explained by the data and not just to other unexplained factors.
that's all fancy words, but the main idea is that a very high R^2 is a good indication that you can estimate data points using this data with a good degree of confidence that you'll be close to what the actual data point would be.