Parsing data from text file to create scatter plot

Hello. If I am posting to the wrong forum please let me know.
Summary:
I have a text file which is attached as "rawdata". It contains records of something (let's call it temperature) at different times on different days.
My goal is to display a graph of temperature versus time so that I can visually analyze trends. I have hundreds of these files, all of different lengths. it is very important that I automate this process as much as possible.

Detail:
(Here I describe what I have done so far; if this is inefficient or unnecessary, feel free to tell me)
I open Excel 2010, click File, Open, and select the file that I want to parse.
It is a TXT file, so the Text Import Wizard comes up. For step one, I select Fixed Width.
I select File Origin: MS-DOS (PC-8).
On step 2 of the wizard, I create column break lines to place all dates in the far left column.
The next column contains the first column of numbers before the first dash (-). The next column contains only the dash - I will later select "ignore this column" to eliminate them. The next column contains the time stamps.
I continue adding column breaks in the wizard until all of the data are parsed into columns in the same manner.

In step 3, I format the first column as "date (DMY)".
The columns with the dashes I select "do not import".
Everything else is "general".
I click "finish", and the resultant workbook is attached, called "import".

Now, as to what I want to do:
I want to display the "temperatures" as a graph vs a date/time axis.
The reason I find this difficult is because the temperatures and times are not in neat columns, but are in 4 columns that go in a left-to-right and top-to-bottom progression and are broken up every few lines.
(I am interested only in numbers that are displayed immediately to the left of a time-stamp. Therefore, the "record #"s should be ignored. We can delete the rows that say "record #" if that helps and can be done automatically.)

Can anyone help me do this?
Thank you so much.