--- loncom/interface/statistics/lonproblemanalysis.pm 2004/11/10 16:56:39 1.103 +++ loncom/interface/statistics/lonproblemanalysis.pm 2004/12/02 21:53:33 1.106 @@ -1,6 +1,6 @@ # The LearningOnline Network with CAPA # -# $Id: lonproblemanalysis.pm,v 1.103 2004/11/10 16:56:39 matthew Exp $ +# $Id: lonproblemanalysis.pm,v 1.106 2004/12/02 21:53:33 matthew Exp $ # # Copyright Michigan State University Board of Trustees # @@ -38,6 +38,7 @@ use Apache::lonstudentsubmissions(); use HTML::Entities(); use Time::Local(); use Spreadsheet::WriteExcel(); +use capa; my $plotcolors = ['#33ff00', '#0033cc', '#990000', '#aaaa66', '#663399', '#ff9933', @@ -193,35 +194,111 @@ sub numerical_response_analysis { $Apache::lonstatistics::enrollment_status, $resource->{'symb'},$respid); # + $problem_analysis->{'answercomputed'} = 1; + if ($problem_analysis->{'answercomputed'}) { + my $answers = + &Apache::lonstathelpers::GetStudentAnswers($r,$problem,$students, + 'Statistics', + 'stats_status'); + $r->print(&numerical_one_dimensional_plot($r,600,150,$answers)); + } + # + if (ref($response_data) ne 'ARRAY') { + $r->print('

'. + &mt('There is no submission data for this resource'). + '

'); + return; + } + my $analysis_html = ''; for (my $plot_num = 1;$plot_num<=$ENV{'form.NumPlots'};$plot_num++) { my $restriction_function; - $restriction_function = sub {($_[0]->{'tries'} == $plot_num?1:0)}; + my $header_message; + my $stats_message; + my $post_message; # passed through &mt sooner rather than later + my $no_data_message; + my @extra_data; + if ($ENV{'form.AnalyzeOver'} eq 'tries') { + $restriction_function = sub {($_[0]->{'tries'} == $plot_num?1:0)}; + $header_message = 'Attempt [_1]'; + $stats_message = + '[_1] submissions, [_2] correct, [_3] incorrect'; + $post_message = ''; + $no_data_message = 'No data exists for attempt [_1]'; + } else { + my $starttime = &Apache::lonhtmlcommon::get_date_from_form + ('startdate_'.$plot_num); + my $endtime = &Apache::lonhtmlcommon::get_date_from_form + ('enddate_'.$plot_num); + ($starttime,$endtime) = &ensure_start_end_times + ($starttime,$endtime, + &get_time_from_row($response_data->[0]), + &get_time_from_row($response_data->[-1]), + $plot_num); + $header_message = 'Data from [_2] to [_3]'; + $extra_data[0] = &Apache::lonlocal::locallocaltime($starttime); + $extra_data[1] = &Apache::lonlocal::locallocaltime($endtime); + # + $stats_message = + '[_1] submissions from [_4] students, [_2] correct, [_3] incorrect'; + # + $post_message = + &mt('Start time: [_1]', + &Apache::lonhtmlcommon::date_setter + ('Statistics','startdate_'.$plot_num,$starttime)). + '
'. + &mt('End time: [_1]', + &Apache::lonhtmlcommon::date_setter + ('Statistics','enddate_'.$plot_num,$endtime)); + $restriction_function = + sub { + my $t = $_[0]->{'timestamp'}; + if ($t >= $starttime && $t < $endtime) { + return 1; + } else { + return 0; + } + }; + $no_data_message = 'No data for [_2] to [_3]'; + } # - my ($correct,$intervals,$answers) = - &numerical_response_determine_intervals($r,$resource,$partid, - $respid,$students); + my ($correct,$answers) = + &numerical_determine_answers($r,$resource,$partid, + $respid,$students); if ($c->aborted()) { return; }; # - my $responses = &classify_response_data($response_data, - $correct, - $restriction_function); - my $student_count = $responses->{'_students'}; - my $correct_count = $responses->{'_correct'}; - my $total_count = $responses->{'_count'}; - my $max = $responses->{'_max'}; - # - my $analysis_html = '

'.&mt('Attempt [_1]',$plot_num).'

'; - $analysis_html.= &numerical_one_dimensional_plot($r,600,150, - scalar(@$students), - $answers); - $analysis_html.= '
'. - ''. - ''. - ''.$/.'
'. - &numerical_plot_percent($r,$responses).''. - &numerical_plot_differences($r,$responses).'
'; - $r->print($analysis_html); + my $responses = &numerical_classify_responses($response_data,$correct, + $restriction_function); + if ($responses->{'_count'} == 0) { + $analysis_html.= + ''. + &mt($no_data_message,$plot_num,@extra_data). + ''; + } else { + $analysis_html.= + ''. + &mt($header_message,$plot_num,@extra_data). + ''. + ''. + &mt($stats_message, + $responses->{'_count'}, + $responses->{'_correct'}, + $responses->{'_count'}-$responses->{'_correct'}, + $responses->{'_students'}, + @extra_data). + ''. + ''.''. + &numerical_plot_percent($r,$responses).''. + ''. + &numerical_plot_differences($r,$responses).''. + ''; + } + if ($post_message ne '') { + $analysis_html .= + ''.$post_message.''; + } } + $analysis_html.=''; + $r->print($analysis_html); # return; } @@ -231,21 +308,21 @@ sub numerical_plot_percent { # my $total = $responses->{'_count'}; return '' if ($total == 0); - my $minbin = 0.5; + my $minbin = 5; while (my ($interval,$submissions) = each(%$responses)) { next if ($interval =~ /^_/); my ($ans,$ans_low,$ans_high) = split(" ",$interval); - my $low_percent = abs(($ans-$ans_low)/$ans); - my $high_percent = abs(($ans_high-$ans)/$ans); + my $low_percent = abs(100*($ans-$ans_low)/$ans); + my $high_percent = abs(100*($ans_high-$ans)/$ans); if ($minbin > $high_percent) { $minbin = $high_percent; } if ($minbin > $low_percent) { $minbin = $low_percent; } - } + } # my @bins; if ($minbin < 1) { - @bins = (0.1, ".5",1,1.5,2,2.5,3,4,5,10,20,50,100); + @bins = ('0.1','0.5','1.0','1.5','2.0','2.5','3.0','4.0','5.0',10,20,50,100); } elsif ($minbin < 2) { - @bins = (0.5,1,1.5,2,2.5,3,4,5,10,20,50,100); + @bins = ('0.5','1.0','1.5','2.0','2.5','3.0','4.0','5.0',10,20,50,100); } elsif ($minbin < 5) { @bins = (1,2,3,4,5,10,25,50,75,100,200); } elsif ($minbin < 10) { @@ -257,53 +334,37 @@ sub numerical_plot_percent { # my @correct; my @incorrect; + my @count; while (my ($interval,$submissions) = each(%$responses)) { next if ($interval =~ /^_/); my ($ans,$ans_low,$ans_high) = split(" ",$interval); - while (my ($submission,$count) = each(%$submissions)) { + while (my ($submission,$counts) = each(%$submissions)) { + my ($correct_count,$incorrect_count) = @$counts; my $scaled_value = abs(($submission-$ans)/$ans); my $bin=0; for ($bin=0;$bin<$#bins;$bin++) { # not <= for a reason last if ($bins[$bin]>$scaled_value); } - if (($submission <= $ans_high) && ($submission >= $ans_low)) { - $correct[$bin]+=$count; - } else { - $incorrect[$bin]+=$count; - } + $correct[$bin]+=$correct_count; + $incorrect[$bin]+=$incorrect_count; + $count[$bin]+=$correct_count+$incorrect_count; } } # + my @plot_correct; + my @plot_incorrect; for (my $i=0;$i<=$#bins;$i++) { - if (! defined( $correct[$i])) { $correct[$i]=0;} - if (! defined($incorrect[$i])) { $incorrect[$i]=0;} - $correct[$i]*=100/$total; - $incorrect[$i]*=100/$total; + $plot_correct[$i] = $correct[$i]*100/$total; + $plot_incorrect[$i] = $incorrect[$i]*100/$total; } my $title = &mt('Distribution by Percent'); my $graph = &Apache::loncommon::DrawBarGraph ($title,'Percent difference from correct','Number of answers', - 100,['#33FF00','#FF3300'],\@labels,\@correct,\@incorrect, + 100,['#33FF00','#FF3300'],\@labels,\@plot_correct,\@plot_incorrect, {xskip=>1}); # - my $table = $graph.''. - ''; - for (my $i=0;$i<=$#bins;$i++) { - my $lownum; - if ($i == 0) { - $lownum = 0; - } else { - $lownum = $bins[$i-1]; - } - my $highnum = $bins[$i]; - $table .= - ''. - ''. - ''. - ''. - ''.''.$/; - } - $table.= '
'.&mt('Bar').''.&mt('Range').'
'.$labels[$i].''.$lownum.' - '.$highnum.'
'.$/; + my $table = $graph.$/. + &numerical_bin_table(\@bins,\@labels,\@incorrect,\@correct,\@count).$/; return $table; } @@ -339,68 +400,158 @@ sub numerical_plot_differences { # my @correct; my @incorrect; + my @count; while (my ($interval,$submissions) = each(%$responses)) { next if ($interval =~ /^_/); my ($ans,$ans_low,$ans_high) = split(" ",$interval); - while (my ($submission,$count) = each(%$submissions)) { + while (my ($submission,$counts) = each(%$submissions)) { + my ($correct_count,$incorrect_count) = @$counts; my $value = abs($submission-$ans); my $bin=0; for ($bin=0;$bin<$#bins;$bin++) { # not <= for a reason last if ($bins[$bin]>$value); } - if (($submission <= $ans_high) && ($submission >= $ans_low)) { - $correct[$bin]+=$count; - } else { - $incorrect[$bin]+=$count; - } + $correct[$bin]+=$correct_count; + $incorrect[$bin]+=$incorrect_count; + $count[$bin]+=$correct_count+$incorrect_count; } } # + my @plot_correct; + my @plot_incorrect; for (my $i=0;$i<=$#bins;$i++) { - if (! defined( $correct[$i])) { $correct[$i]=0;} - if (! defined($incorrect[$i])) { $incorrect[$i]=0;} - $correct[$i]*=100/$total; - $incorrect[$i]*=100/$total; + $plot_correct[$i] = $correct[$i]*100/$total; + $plot_incorrect[$i] = $incorrect[$i]*100/$total; } my $title = &mt('Distribution by Magnitude'); my $graph = &Apache::loncommon::DrawBarGraph ($title,'magnitude difference from correct','Number of answers', - 100,['#33FF00','#FF3300'],\@labels,\@correct,\@incorrect,{xskip=>1}); + 100,['#33FF00','#FF3300'],\@labels,\@plot_correct,\@plot_incorrect, + {xskip=>1}); # - my $table = $graph.''. - ''; - for (my $i=0;$i<=$#bins;$i++) { + my $table = $graph.$/. + &numerical_bin_table(\@bins,\@labels,\@incorrect,\@correct,\@count).$/; + return $table; +} + +sub numerical_classify_responses { + &Apache::lonnet::logthis('--------------'); + my ($full_row_data,$correct,$function) = @_; + my %submission_data; + my %students; + my $max=0; + foreach my $row (@$full_row_data) { +# &Apache::lonnet::logthis(' row = '.join(',',@$row)); + my %subm = &hashify_attempt($row); + if (ref($correct) eq 'HASH') { + $subm{'correct'} = $correct->{$subm{'student'}}->{'answer'}; + $subm{'unit'} = $correct->{$subm{'student'}}->{'unit'}; + } else { # This probably never happens.... + $subm{'correct'} = $correct->{'answer'}; + $subm{'unit'} = $correct->{'unit'}; + } + $subm{'submission'} =~ s/=\d+\s*$//; + if (&$function(\%subm)) { + my $scaled = '1'; + my ($sname,$sdom) = split(':',$subm{'student'}); + # Note that $subm{'unit'} is modified by the following call + # We do not use it again but you should be aware just in case. +# my ($myunit,$mysub) = ($subm{'unit'},$subm{'submission'}); +# $myunit = 'm'; +# # &Apache::lonnet::logthis($myunit); +# my $result = +# &capa::caparesponse_get_real_response($myunit, +# $mysub, +# \$scaled); +# # &Apache::lonnet::logthis(' '.$myunit.':'.$subm{'unit'}. +# # ' '.$mysub.':'.$subm{'submission'}.'; '.$result); +# next if (! defined($scaled)); +# next if ($result ne '6'); +# my $submission = $scaled; + my $submission = $subm{'submission'}; + $students{$subm{'student'}}++; + if (&numerical_submission_is_correct($subm{'award'})) { + &Apache::lonnet::logthis('correct:'.$submission.':'.$subm{'correct'}); + $submission_data{'_correct'}++; + $submission_data{'_count'}++; + $submission_data{$subm{'correct'}}->{$submission}->[0]++; + } elsif (&numerical_submission_is_incorrect($subm{'award'})) { + &Apache::lonnet::logthis('incorrect:'.$submission.':'.$subm{'correct'}); + $submission_data{'_count'}++; + $submission_data{$subm{'correct'}}->{$submission}->[1]++; + } + my $value = + $submission_data{$subm{'correct'}}->{$submission}->[0]+ + $submission_data{$subm{'correct'}}->{$submission}->[1]; + if ($max < $value) { $max = $value; } + } + } + $submission_data{'_max'} = $max; + $submission_data{'_students'}=scalar(keys(%students)); + return \%submission_data; +} + +sub numerical_submission_is_correct { + my ($award) = @_; + &Apache::lonnet::logthis('award = "'.$award.'"'); + if ($award =~ /^(APPROX_ANS|EXACT_ANS)$/) { + return 1; + } else { + return 0; + } +} + +sub numerical_submission_is_incorrect { + my ($award) = @_; + if ($award =~ /^(INCORRECT)$/) { + return 1; + } else { + return 0; + } +} + +sub numerical_bin_table { + my ($bins,$labels,$incorrect,$correct,$count)=@_; + my $table = + '
'.&mt('Bar').''.&mt('Range').'
'. + ''. + ''. + ''. + ''. + ''.$/; + for (my $i=0;$i[$i-1]; } - my $highnum = $bins[$i]; + my $highnum = $bins->[$i]; $table .= ''. - ''. + ''. ''. ''. - ''.''.$/; + ''. + ''. + ''. + ''. + ''.$/; } - $table.= '
'.&mt('Bar').''.&mt('Range').''.&mt('Incorrect').''.&mt('Correct').''.&mt('Count').'
'.$labels[$i].''.$labels->[$i].''.$lownum.' - '.$highnum.'
'.$highnum.''.$incorrect->[$i].''.$correct->[$i].''.$count->[$i].'
'.$/; + $table.= ''; return $table; } -sub numerical_response_determine_intervals { +sub numerical_determine_answers { my ($r,$resource,$partid,$respid,$students)=@_; my $c = $r->connection(); # - # FIX ME: Need progress dialog updates + # FIX ME: May need progress dialog updates # # Read in the cache (if it exists) before we start timing things. &Apache::lonstathelpers::ensure_proper_cache($resource->{'symb'}); # my $correct; - # %intervals differs from %answers because it may be possible for two - # students to have the same correct answer but different intervals. - my %intervals; my %answers; foreach my $student (@$students) { last if ($c->aborted()); @@ -413,25 +564,30 @@ sub numerical_response_determine_interva $sdom); # make the key my $key = $partid.'.'.$respid; - my $interval = join(' ',($analysis->{$key.'.answer'}->[0], - $analysis->{$key.'.ans_low'}->[0], - $analysis->{$key.'.ans_high'}->[0])); - $correct->{$sname.':'.$sdom} = $interval; - $intervals{$interval}++; + $correct->{$sname.':'.$sdom}->{'answer'} = + $analysis->{$key.'.answer'}->[0]; + $correct->{$sname.':'.$sdom}->{'unit'} = + $analysis->{$key.'.unit'}->[0]; $answers{$analysis->{$key.'.answer'}->[0]}++; } &Apache::lonstathelpers::write_analysis_cache(); - return ($correct,\%intervals,\%answers); + return ($correct,\%answers); } +# +# Inputs: $r, $width, $height, $data +# $n = number of students +# $data = hashref of $answer => $frequency pairs sub numerical_one_dimensional_plot { - my ($r,$width,$height,$n,$data)=@_; + my ($r,$width,$height,$data)=@_; # # Compute data -> image scaling factors my $max_y = 0; my $min_x = undef; my $max_x = undef; + my $n = 0; while (my ($answer,$count) = each(%$data)) { + $n+=$count; $max_y = $count if ($max_y < $count); if (! defined($min_x) || $answer < $min_x) { $min_x = $answer; @@ -441,7 +597,11 @@ sub numerical_one_dimensional_plot { } } # - my $h_scale = ($width-10)/($max_x-$min_x); + my $min_max_difference = $max_x - $min_x; + if (! defined($min_max_difference) || $min_max_difference == 0) { + $min_max_difference = 1; + } + my $h_scale = ($width-10)/$min_max_difference; # my $ticscale = 5; if ($max_y * $ticscale > $height/2) { @@ -471,7 +631,6 @@ sub numerical_one_dimensional_plot { } $plot .= ''; my $plotresult = &Apache::lonxml::xmlparse($r,'web',$plot); - my $result = ''. '
'. ''.&mt('Distribution of correct answers').''. @@ -522,33 +681,6 @@ sub circle { return &arc($x,$y,$radius,$radius,0,360,$color,$thickness,$filled); } -sub build_student_data_worksheet { - my ($workbook,$format) = @_; - my $rows_output = 3; - my $cols_output = 0; - my $worksheet = $workbook->addworksheet('Student Data'); - $worksheet->write($rows_output++,0,'Student Data',$format->{'h3'}); - my @Headers = ('full name','username','domain','section', - "student\nnumber",'identifier'); - $worksheet->write_row($rows_output++,0,\@Headers,$format->{'header'}); - my @Students = @Apache::lonstatistics::Students; - my $studentrows = &Apache::loncoursedata::get_student_data(\@Students); - my %ids; - foreach my $row (@$studentrows) { - my ($mysqlid,$student) = @$row; - $ids{$student}=$mysqlid; - } - foreach my $student (@Students) { - my $name_domain = $student->{'username'}.':'.$student->{'domain'}; - $worksheet->write_row($rows_output++,0, - [$student->{'fullname'}, - $student->{'username'},$student->{'domain'}, - $student->{'section'},$student->{'id'}, - $ids{$name_domain}]); - } - return $worksheet; -} - ######################################################### ######################################################### ## @@ -570,6 +702,9 @@ sub radio_response_analysis { my $analysis_html; my $foildata = $problem_analysis->{'_Foils'}; my ($table,$foils,$concepts) = &build_foil_index($problem_analysis); + if (! defined($concepts)) { + $concepts = []; + } # my %true_foils; my $num_true = 0; @@ -965,7 +1100,6 @@ sub classify_response_data { if (&$function(\%subm)) { $students{$subm{'student'}}++; $submission_data{'_count'}++; - if (&submission_is_correct($subm{'award'})) { $submission_data{'_correct'}++; } @@ -1129,7 +1263,7 @@ sub OR_tries_analysis { # $concept_graph = &Apache::loncommon::DrawBarGraph ('Correct Concepts','Concept Number','Percent Correct', - 100,$plotcolors,undef,\@concept_plot_data); + 100,$plotcolors,undef,\@concept_plot_data,{xskip=>1}); } # # Create Foil Plots @@ -1167,7 +1301,7 @@ sub OR_tries_analysis { # my $correct_graph = &Apache::loncommon::DrawBarGraph ('Correct Statements','Statement','% Answered Correct', - 100,$plotcolors,\@Labels,$Datasets[0]); + 100,$plotcolors,\@Labels,$Datasets[0],{xskip=>1}); # # @@ -1179,7 +1313,7 @@ sub OR_tries_analysis { $response_data{'_correct'}->[$try]; my $incorrect_graph = &Apache::loncommon::DrawBarGraph ('Incorrect Statements','Statement','% Chosen Incorrectly', - 100,$plotcolors,\@Labels,@Datasets); + 100,$plotcolors,\@Labels,@Datasets,{xskip=>1}); $analysis_html.= '
'. ''. @@ -1373,7 +1507,8 @@ sub OR_Foil_Time_Analysis { 100, $plotcolors, undef, - $plotdata[0]); + $plotdata[0], + {xskip=>1}); for (my $j=0; $j< scalar(@{$plotdata[0]});$j++) { $plotdata[0]->[$j]=0; } @@ -1384,7 +1519,7 @@ sub OR_Foil_Time_Analysis { 100, $plotcolors, undef, - @plotdata); + @plotdata,{xskip=>1}); return ($correct_plot,$incorrect_plot); } @@ -1415,7 +1550,7 @@ sub OR_Concept_Time_Analysis { 100, $plotcolors, undef, - \@plotdata); + \@plotdata,{xskip=>1}); } sub OR_time_process_data { @@ -1463,6 +1598,33 @@ sub OR_time_process_data { ## ######################################################### ######################################################### +sub build_student_data_worksheet { + my ($workbook,$format) = @_; + my $rows_output = 3; + my $cols_output = 0; + my $worksheet = $workbook->addworksheet('Student Data'); + $worksheet->write($rows_output++,0,'Student Data',$format->{'h3'}); + my @Headers = ('full name','username','domain','section', + "student\nnumber",'identifier'); + $worksheet->write_row($rows_output++,0,\@Headers,$format->{'header'}); + my @Students = @Apache::lonstatistics::Students; + my $studentrows = &Apache::loncoursedata::get_student_data(\@Students); + my %ids; + foreach my $row (@$studentrows) { + my ($mysqlid,$student) = @$row; + $ids{$student}=$mysqlid; + } + foreach my $student (@Students) { + my $name_domain = $student->{'username'}.':'.$student->{'domain'}; + $worksheet->write_row($rows_output++,0, + [$student->{'fullname'}, + $student->{'username'},$student->{'domain'}, + $student->{'section'},$student->{'id'}, + $ids{$name_domain}]); + } + return $worksheet; +} + sub OR_excel_sheet { my ($r,$resource,$performance_data,$ORdata) = @_; my $response = '';