I can. If they are stored in an SQL database then it will be much faster and easier to when it comes to producing the graphs.
Over time the text files will become large and have a lot of data to sort through to just find the SNRM values. Look at that one row of data I posted and you will see how many values are in the text file... but you are only searching for one figure out of that whole row. If stored as text then your program will have to loop through 100's of 1000's of data to find the one figure in each row that it is looking for.
If you create a database that stores all the SNR values in one table, then you can do a direct query on that table and ask it to pull all the data between 'x & y' date and then send those figures to the google chart code. More code at the beginning to set up all the tables and to create a script which will add the incoming data from DSLstats to your database. But the graph execution and webpage load will be a heck of a lot quicker.
I don't know if you've ever used HG612stats GUI to produce graphs, but it can take a few mins to generate the graphs, that's because all the data is stored as text files. Even on a fast PC it can still take a while.. it wouldn't be the best approach if you're running it from a remote web-site.
You need to sit down and do some serious thinking how you are going to be outputting the data... and which method is going to be best depending on how you are going to output that data. What may work best on a machine running a local web-server, or something to just generate static graphs is going to have different requirements to running it on a remote website. It also depends on how often you think you are going to view them.
---
There is really an awful lot of thought that has gone behind what Tony did, its not something that will have taken him a couple of days. Unless you have to do this sort of thing, then people don't realise how much work goes into being able to present the data.
To give an example... my broadband checker gets all the data from the BTw database I do a simple query and it returns all the raw data in xml format. I immediately have the info that is needed.
The bit that took me weeks/months to code is being able to break down the raw code and then output the data so that its readable to humans and looks pretty. With MDWS, google graphs is doing the pretty bit, but there is still an awful lot of code in the middle that will be needed to make sure that the graphs get the right info they need.