Filtering Confusion!

Hi

I have created a matrix report in Power Bi Desktop to display student weekly attendance for the last 5 weeks, so setup is as follows: Rows - Student Name,

Columns - Week Commencing,

Values - % Attendance (a simple measure %Attendance = DIVIDE(SUM('Weekly Attendance'[PRESENT_MARKS]),SUM('Weekly Attendance'[ALL_MARKS]))

The basic matrix works fine but I want to only show data for students that have less than 85% Attendance for the last 2 weeks with the added complication that I only want to include weeks when the school is open.  So I have written a little SQL that aggregates the student attendance by week & course (I'm using direct query for the matrix) and created 2 extra columns that show the Sum of the present marks & all marks just for the last 2 weeks (excluding the holidays).  I have created a measure exactly the same as the one above but for the columns with the last 2 weeks attendance & tried to use that as a visual filter to display only those students with less than 85% attendance.  However, this does not work and produces the error 'Couldn't load the data for this visual - Conversion of a nvarchar data type to a datetime data type resulted in an out-of-range value'.  The error message doesn't make any sense but should I able to filter like this or am I tackling this completely the wrong way?  Would be very helpful if someone could explain how I could best filter the visual for this scenario.

Data from query looks like this:

Student Name  Week Commencing   Present Marks   All Marks  PresentMarksLast2Wks  AllMarksLast2Wks

Fred                  05-12-2016                10                      12             10                                    12

Fred                  12-12-2016                10                      13             10                                    13

Fred                  28-11-2016                11                       18             0                                      0