A good way to combine three graphs with the same Y axis but different X axis. I have three graphs as show below (the last one is split because I couldn’t take a screenshot with everything together) 

What’s a good approach (from a UX standpoint) to combining all three graphs into a single visualization? What kind of visualization should I be looking at? 
 A: I don't think you should try to combine these. The year, the day of the month, and the day of the week are all orthogonal concerns. If you layered them into a single plot with a three-layer stack of $x$-labels or something, you'd just have a more cluttered presentation.
Here's some more miscellaneous advice:


*

*Don't cumulate. The cumulative display stretches the $y$-axis and makes it hard to make comparisons across $x$-values, since the reader needs to eyeball the difference on the $y$-axis rather than being able to read off the $y$-axis directly.

*Remove the fill under the lines, and remove the vertical gridlines. These add a lot of visual noise for little information.

*Present $y$-values as per-day means rather than sums. This will make the numbers smaller and easier to interpret, and it will let you put the $y$-axes of the three figures on the same scale.

*Add thousands separators to the $y$-labels to make them more legible.


I recommend the popular book The Visual Display of Quantitative Information, by Edward R. Tufte, for guidance on making easy-to-read data graphics.
A: Here below you will find some suggestions, each one depending on the purpose and on the prior knowledge of the users.
Just show the actual data and let the user pick up from there
Plot the actual number of visits per period (day or month), instead of the cumulated data, and use line as an encoding mark. In terms of granularity, you can continue to stick with the multiple approach, one for each zoom level, or alternatively, one for each aggregation measure (total, average, min, max, etc.).
Let the user configure the period
Interactive displays offer the advantage that the user can zoom in and out or filter on the go to focus on a smaller period. Here is one example that pretty much speaks for itself: overview+detail based on vega.js
Reveal the underlying trend to the user
A convenient and okay-to-understand approach is to calculate moving averages and plot those as a layer over the original daily or monthly data. Based on the number of data points involved in the moving average calculation, you can visualise long or short term trends.
At what time of the day/week/month/year does the user's site have most visits?
One way to answer this question is to draw the data on a heatmap with a time scale on the x axis, e.g. months (or day numbers), and an ordinal scale on the y axis, e.g. day number (or even intra-day time). The number of visits should be encoded with grey tones (brightness) on a continuous scale. Here is such a heatmap based on vega.js.
How do the site's daily visits compare to the prior year average?
A simple way of comparing data in time is to transform the daily or monthly data into index values, base 100 being e.g the average number of daily or monthly visits during the previous year. Obviously, the choice of the base 100 requires careful consideration: it should be easily understood by the users on one hand and be meaningful on the other.
