No, i think it is not sufficient to find out regression analysis in above figure earlier you have to mentioned .

I am not familiar with your application. The fact that your x-axis is restricted to 0 to 100% may have something to do with your regression not looking linear for the first 6% or so, and maybe the last 2%. In the linear looking range, linear regression may be good. I expect you might see heteroscedasticity, which is natural. You could do a graphical residual analysis, plotting predicted y values on the x-axis, and estimated residuals on the y-axis.

Cheers!