How do I find the corresponding value for a given data and graph in Excel?

I agree with.  I don't know how to do what you're asking about without using an equation, but the reason you cite for not allowing equations I don't think makes sense.  You said that you need to plot the y = 50 point as the reason for not using an equation, but I think you can plot it and have the corresponding x value generated by an equation. 1. Plot data using a scatter plot 2. add a "trendline", y=96.406e^-0.221x fits the data with R^2 = 0.998, or a parabola fits the data almost as well: 1.0242x^2 - 18.812x + 98,114 has R^2 - 0.997. 3. decide on your fit equation and then add a row at the bottom where you can enter the y value and solve for x.  You'll need to solve for x in your fit equation though, or simply conceive of x as y and y as x so that Excel gives you your fit line speaking in terms of x in the first place.  Make sure that your plot range includes the new row you add so that it will be plotted, or you can plot it in a new range so it won't affect the fit line, but will show up as another color on the plot.

You do need to use an equation but Excel has an easy function to do that. Checkout LINEST. Basically you need to use Linest between the two points based on where your question point is. To plot it on the same graph is easy, it's just how you set up your graph range.

What you are really asking for is not to read data from the chart, but to interpolate values from the data you originally put into it. In the data you provided there simply is no data pair (x,50), so your chart will never be able to return x without doing some kind of interpolation between the two surrounding points (2.5,55) and (5,32).Of course Excel shows  lines between the points in your chart, and that suggest that there is data, but there really isn't. The lines are simply a linear or some kind of polynomal interpolation between the surrounding points.So the only right way to do this is by adding a trend line. Select your data points and choose insert trend line. It is up to you to decide whether it will be a linear trend line, a logarithmic or a 3rd order polynomal. Make sure you check the "show equation" box. Then Excel will add a trendline and show its equation. Using that equation you can simply calculate X from any given Y. As far as I know the only way to do this is type the equation manually in your workbook. Keep in mind however that any trend line is always an approximation, and that you better not use it to extrapolate. Below two screenshots (in Dutch).

On a scatter plot, you can see the value by clicking on the data point; it is still awkward because Excell displays the row number instead of the actual x coordinate. 

Edited to add: WOW! When I started typing about a half hour ago, there were no answers and I was feeling real cleaver. Then when I submitted my answer, there were 8 others already there. There's a lot of smart (and fast) people on Quora. It sounds like what you're looking for is regression analysis. For instance, I want to predict how tall someone was from the length of his or her femur. I measure the femur length and total height of a whole bunch of skeletons. I then put the data in a table and draw a scatter chart from that, where femur length is the x-axis and height is the y-axis. Note: In real life, I may want to narrow the population, such as sex, geographic ancestry, age, etc. So now you've got a bunch of points that has a rough pattern showing the longer the femur, the taller the height (duh). Performing some statistical magic, I find the best fit line through the points for a white female is the equation: Height = 2.47 * femur length + 54.10 plus or minus 3.72 cm. I  plug in a known x (femur length) to get the value of y. What you're wanting is pretty much the same thing. In this case, you either use basic algebra to solve y for a known x, or just swap your axes. The trick is the statistical magic that comes up with the equation. Luckily, Excel has a quick and dirty method to do this. After you create your chart, you want to insert a trendline. With your trendline, you'll want Excel to include: 1) The equation for the trendline . 2) The R² value. This is the coefficient of determination.  This is a number that indicates how well data fit a statistical model – sometimes simply a line or a curve. An R² of 1 indicates that the regression line perfectly fits the data, while an R² of 0 indicates that the line does not fit the data at all. Excel allows you to insert several different kinds of trendlines: 1) Exponential 2) Linear 3) Logarithmic 4) Polynomial 5) Power 6) Moving Average (this is probably not the trend line you're interested in). What I would do is insert a trendline, note its R² value, delete it, add a different type of trend line, note its R² value, and so on for each type. Then I would use whichever equation had the R² value closest to 1. Note: It is up to you to determine if the R² is high enough or not. It's entirely possible that no useful equation can be determined from the data if it has little or no pattern. To insert a trendline into a chart: 1) Select the chart by clicking on it. 2) You'll now see the Chart Tools on the ribbon with the tabs "Design," "Layout," and "Format." Click on the "Layout" tab. 3) In the "Analysis" group is a button for "Trendline." Click the drop down arrow and choose "More Trendline Options . . . ". 4) This will open the "Format Trendline" dialog box.      a) Select a trendline type.      b) At the very bottom, click the checkboxes for "Display equation on chart" and "Display R² value on chart." 5) Click the "Close" button. You've now got a trendline. Note: This is NOT statistically rigorous, but, like I said, may be a good enough method to get you want you want.