--- loncom/interface/statistics/lonproblemanalysis.pm 2005/02/16 17:43:20 1.111 +++ loncom/interface/statistics/lonproblemanalysis.pm 2005/03/15 00:51:43 1.119 @@ -1,6 +1,6 @@ # The LearningOnline Network with CAPA # -# $Id: lonproblemanalysis.pm,v 1.111 2005/02/16 17:43:20 matthew Exp $ +# $Id: lonproblemanalysis.pm,v 1.119 2005/03/15 00:51:43 matthew Exp $ # # Copyright Michigan State University Board of Trustees # @@ -97,6 +97,9 @@ sub BuildProblemAnalysisPage { $r->print($html.(' 'x5)); } # + $r->print(&Apache::lonstathelpers::submission_report_form + ('problem_analysis')); + # $r->print('
'); $r->rflush(); # @@ -104,7 +107,7 @@ sub BuildProblemAnalysisPage { my $current_problem = &Apache::lonstathelpers::get_target_from_id ($ENV{'form.problemchoice'}); # - my ($prev,$curr,$next) = + my ($navmap,$prev,$curr,$next) = &Apache::lonstathelpers::get_prev_curr_next($current_problem, $problem_types, 'response', @@ -127,14 +130,15 @@ sub BuildProblemAnalysisPage { $r->print('resource is undefined'); } else { my $resource = $current_problem->{'resource'}; - $r->print('

'.$resource->{'title'}.'

'); - $r->print('

'.$resource->{'src'}.'

'); + $r->print('

'.$resource->compTitle.'

'); + $r->print('

'.$resource->src.'

'); + $r->print('

'.&Apache::lonstatistics::section_and_enrollment_description().'

'); if ($ENV{'form.show_prob'} eq 'true') { $r->print(&Apache::lonstathelpers::render_resource($resource)); } $r->rflush(); my %Data = &Apache::lonstathelpers::get_problem_data - ($resource->{'src'}); + ($resource->src); my $problem_data = $Data{$current_problem->{'part'}. '.'. $current_problem->{'respid'}}; @@ -159,8 +163,7 @@ sub BuildProblemAnalysisPage { &mt('Analyze Problem').'" />'); $r->print(' 'x5); $r->print('

'.&mt('Please select a problem to analyze').'

'); - $r->print(&Apache::lonstathelpers::ProblemSelector - ($problem_types)); + $r->print(&Apache::lonstathelpers::problem_selector($problem_types)); } } @@ -184,9 +187,9 @@ sub numerical_response_analysis { $problem->{'respid'}); # Gather student data my $response_data = &Apache::loncoursedata::get_response_data - (\@Apache::lonstatistics::SelectedSections, + ([&Apache::lonstatistics::get_selected_sections()], $Apache::lonstatistics::enrollment_status, - $resource->{'symb'},$respid); + $resource->symb,$respid); # $problem_analysis->{'answercomputed'} = 1; if ($problem_analysis->{'answercomputed'}) { @@ -260,9 +263,10 @@ sub numerical_response_analysis { $respid,$students); if ($c->aborted()) { return; }; # - my $responses = &numerical_classify_responses($response_data,$correct, - $restriction_function); - if ($responses->{'_count'} == 0) { + my ($responses,$stats) = + &numerical_classify_responses($response_data,$correct, + $restriction_function); + if ($stats->{'submission_count'} == 0) { $analysis_html.= ''. &mt($no_data_message,$plot_num,@extra_data). @@ -274,16 +278,16 @@ sub numerical_response_analysis { ''. ''. &mt($stats_message, - $responses->{'_count'}, - $responses->{'_correct'}, - $responses->{'_count'}-$responses->{'_correct'}, - $responses->{'_students'}, + $stats->{'submission_count'}, + $stats->{'correct_count'}, + $stats->{'incorrect_count'}, + $stats->{'students'}, @extra_data). ''. ''.''. - &numerical_plot_percent($r,$responses).''. + &numerical_plot_percent($r,$responses,$stats).''. ''. - &numerical_plot_differences($r,$responses).''. + &numerical_plot_differences($r,$responses,$stats).''. ''; } if ($post_message ne '') { @@ -298,63 +302,76 @@ sub numerical_response_analysis { } sub numerical_plot_percent { - my ($r,$responses) = @_; + my ($r,$responses,$stats) = @_; # - my $total = $responses->{'_count'}; + my $total = $stats->{'submission_count'}; return '' if ($total == 0); - my $minbin = 5; - while (my ($interval,$submissions) = each(%$responses)) { - next if ($interval =~ /^_/); - my ($ans,$ans_low,$ans_high) = split(" ",$interval); - 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 $max_bins = 50; + my $lowest_percent = $stats->{'min_percent'}; + my $highest_percent = $stats->{'max_percent'}; + my $percent_spread = $highest_percent - $lowest_percent; + foreach (qw/20 30 40 50 100 200/) { + if ($percent_spread < $_) { + $highest_percent =$_/2; + last; + } + } + my $bin_size = 1; + foreach (qw/0.01 0.05 0.1 0.5 1 2 5 10 20 25 50 100/) { + if ($lowest_percent/2 < $_){ + $bin_size = $_; + last; + } } - # my @bins; - if ($minbin < 1) { - @bins = ('0.1','0.5','1.0','1.5','2.0','2.5','3.0','4.0','5.0',10,20,50,100); - } elsif ($minbin < 2) { - @bins = ('0.5','1.0','1.5','2.0','2.5','3.0','4.0','5.0',10,20,50,100); - } elsif ($minbin < 5) { - @bins = (1,2,3,4,5,10,25,50,75,100,200); - } elsif ($minbin < 10) { - @bins = (2,4,6,8,10,12,15,20,25,30,50,75,100,200); - } else { - @bins = (5,10,15,20,25,30,50,75,100,200); + for (my $bin = -$highest_percent;$bin<0;$bin+=$bin_size) { + push (@bins,$bin); } - my @labels = (1..scalar(@bins)); + for (my $bin = 0; $bin<$highest_percent;$bin+=$bin_size) { + push (@bins,$bin); + } + push(@bins,$highest_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 ($ans,$submissions) = each(%$responses)) { while (my ($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 + my $scaled_value = 100*($submission-$ans)/abs($ans); + if ($scaled_value < $bins[0]) { + $bins[0]=$scaled_value -1; + } + my $bin; + for ($bin=0;$bin<$#bins;$bin++) { last if ($bins[$bin]>$scaled_value); } - $correct[$bin]+=$correct_count; - $incorrect[$bin]+=$incorrect_count; - $count[$bin]+=$correct_count+$incorrect_count; + $correct[$bin-1]+=$correct_count; + $incorrect[$bin-1]+=$incorrect_count; + $count[$bin-1]+=$correct_count+$incorrect_count; } } # - my @plot_correct; - my @plot_incorrect; - for (my $i=0;$i<=$#bins;$i++) { - $plot_correct[$i] = $correct[$i]*100/$total; - $plot_incorrect[$i] = $incorrect[$i]*100/$total; + my @plot_correct = @correct; + my @plot_incorrect = @incorrect; + my $max; + for (my $i=0;$i<$#bins;$i++) { + $plot_correct[$i] *= 100/$total; + $plot_incorrect[$i] *= 100/$total; + if (! defined($max) || + $max < $plot_correct[$i]+$plot_incorrect[$i] ) { + $max = $plot_correct[$i]+$plot_incorrect[$i]; + } } - my $title = &mt('Distribution by Percent'); + foreach (qw/1 5 10 15 20 25 30 40 50 75 100/) { + if ($max <$_) { $max = $_; last; } + } + # + my $title = &mt('Percent Difference'); + my @labels = (1..scalar(@bins)-1); my $graph = &Apache::loncommon::DrawBarGraph - ($title,'Percent difference from correct','Number of answers', - 100,['#33FF00','#FF3300'],\@labels,\@plot_correct,\@plot_incorrect, + ($title,'Percent Difference from Correct','Percent of Answers', + $max,['#33FF00','#FF3300'],\@labels,\@plot_correct,\@plot_incorrect, {xskip=>1}); # my $table = $graph.$/. @@ -363,64 +380,66 @@ sub numerical_plot_percent { } sub numerical_plot_differences { - my ($r,$responses) = @_; + my ($r,$responses,$stats) = @_; # - my $total = $responses->{'_count'}; + my $total = $stats->{'submission_count'}; return '' if ($total == 0); - my $minbin = undef; - my $maxbin = undef; - while (my ($interval,$submissions) = each(%$responses)) { - next if ($interval =~ /^_/); - my ($ans,$ans_low,$ans_high) = split(" ",$interval); - my $low_diff = abs($ans-$ans_low); - my $high_diff = abs($ans_high-$ans); - if (! defined($maxbin)) { $maxbin = $low_diff;} - if (! defined($minbin)) { $minbin = $low_diff;} - # - if ($minbin > $high_diff) { $minbin = $high_diff; } - if ($minbin > $low_diff ) { $minbin = $low_diff; } - # - if ($maxbin < $high_diff) { $maxbin = $high_diff; } - if ($maxbin < $low_diff ) { $maxbin = $low_diff; } - } - # + my $max_bins = 21; + my $min_bin_size = $stats->{'min_abs'}; + my $low_bin = $stats->{'lowest_ans'}-$stats->{'max_bin_size'}; + my $high_bin = $stats->{'highest_ans'}+$stats->{'max_bin_size'}; + if ($high_bin > 0 && $low_bin > -$high_bin) { + $low_bin = -$high_bin; + } elsif ($low_bin < 0 && $high_bin < -$low_bin) { + $high_bin = -$low_bin; + } + if (($high_bin -$low_bin)/$min_bin_size * 2 > $max_bins) { + $min_bin_size = abs($high_bin - $low_bin) / $max_bins * 2; + } my @bins; - my @labels; - # Hmmmm, should switch to absolute difference - for (my $i=1;$i<=20;$i++) { - push(@bins,$i*$minbin/2); - push(@labels,$i); + for (my $num = $low_bin;$num <= $high_bin;$num+=($min_bin_size/2)) { + push(@bins,$num); } # 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 ($ans,$submissions) = each(%$responses)) { while (my ($submission,$counts) = each(%$submissions)) { my ($correct_count,$incorrect_count) = @$counts; - my $value = abs($submission-$ans); + my $scaled_value = $submission-$ans; + if ($scaled_value < $bins[0]) { + $bins[0]=$scaled_value-1; + } my $bin=0; - for ($bin=0;$bin<$#bins;$bin++) { # not <= for a reason - last if ($bins[$bin]>$value); + for ($bin=0;$bin<$#bins;$bin++) { + last if ($bins[$bin]>$scaled_value); } - $correct[$bin]+=$correct_count; - $incorrect[$bin]+=$incorrect_count; - $count[$bin]+=$correct_count+$incorrect_count; + $correct[$bin-1]+=$correct_count; + $incorrect[$bin-1]+=$incorrect_count; + $count[$bin-1]+=$correct_count+$incorrect_count; } } - # - my @plot_correct; - my @plot_incorrect; + my @plot_correct = @correct; + my @plot_incorrect = @incorrect; + my $max; for (my $i=0;$i<=$#bins;$i++) { - $plot_correct[$i] = $correct[$i]*100/$total; - $plot_incorrect[$i] = $incorrect[$i]*100/$total; + $plot_correct[$i] *= 100/$total; + $plot_incorrect[$i] *= 100/$total; + if (! defined($max) || + $max < $plot_correct[$i]+$plot_incorrect[$i] ) { + $max = $plot_correct[$i]+$plot_incorrect[$i]; + } } - my $title = &mt('Distribution by Magnitude'); + foreach (qw/1 5 10 15 20 25 30 40 50 75 100/) { + if ($max <$_) { $max = $_; last; } + } + # + my $title = &mt('Difference between submission and correct'); + my @labels = (1..scalar(@bins)-1); my $graph = &Apache::loncommon::DrawBarGraph - ($title,'magnitude difference from correct','Number of answers', - 100,['#33FF00','#FF3300'],\@labels,\@plot_correct,\@plot_incorrect, + ($title,'Difference from Correct','Percent of Answers', + $max,['#33FF00','#FF3300'],\@labels,\@plot_correct,\@plot_incorrect, {xskip=>1}); # my $table = $graph.$/. @@ -429,60 +448,101 @@ sub numerical_plot_differences { } sub numerical_classify_responses { - &Apache::lonnet::logthis('--------------'); my ($full_row_data,$correct,$function) = @_; my %submission_data; my %students; + my %stats; 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'}; + my $s_correct = $correct->{$subm{'student'}}; + $subm{'correct'} = $s_correct->{'answer'}; + foreach my $item ('unit','ans_low','ans_high') { + $subm{$item} = $s_correct->{$item}; + } } else { # This probably never happens.... $subm{'correct'} = $correct->{'answer'}; $subm{'unit'} = $correct->{'unit'}; } + # + my $abs_low =abs($subm{'correct'}-$subm{'ans_low'}); + my $abs_high=abs($subm{'correct'}-$subm{'ans_high'}); + if (! defined($stats{'min_abs'}) || + $stats{'min_abs'} > $abs_low) { + $stats{'min_abs'} = $abs_low; + } + if ($stats{'min_abs'} > $abs_high) { + $stats{'min_abs'} = $abs_high; + } + if (! defined($stats{'max_abs'}) || + $stats{'max_abs'} < $abs_low) { + $stats{'max_abs'} = $abs_low; + } + if ($stats{'max_abs'} < $abs_high) { + $stats{'max_abs'} = $abs_high; + } + my $low_percent; + my $high_percent; + if (defined($subm{'correct'}) && $subm{'correct'} != 0) { + $low_percent = 100 * abs($abs_low / $subm{'correct'}); + $high_percent = 100 * abs($abs_high / $subm{'correct'}); + } + if (! defined($stats{'min_percent'}) || + $stats{'min_percent'} > $low_percent) { + $stats{'min_percent'} = $low_percent; + } + if ($stats{'min_percent'} > $high_percent) { + $stats{'min_percent'} = $high_percent; + } + if (! defined($stats{'max_percent'}) || + $stats{'max_percent'} < $low_percent) { + $stats{'max_percent'} = $low_percent; + } + if ($stats{'max_percent'} < $high_percent) { + $stats{'max_percent'} = $high_percent; + } + if (! defined($stats{'lowest_ans'}) || + $stats{'lowest_ans'} > $subm{'correct'}) { + $stats{'lowest_ans'} = $subm{'correct'}; + } + if (! defined($stats{'highest_ans'}) || + $stats{'highest_ans'} < $subm{'correct'}) { + $stats{'highest_ans'} = $subm{'correct'}; + } + # $subm{'submission'} =~ s/=\d+\s*$//; 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'}); my $result = &capa::caparesponse_get_real_response($myunit, $mysub, \$scaled); +# &Apache::lonnet::logthis('scaled = '.$scaled.' result ='.$result); next if (! defined($scaled)); - next if ($result ne '6'); +# next if ($result ne '6'); my $submission = $scaled; $students{$subm{'student'}}++; + $stats{'submission_count'}++; if (&numerical_submission_is_correct($subm{'award'})) { - &Apache::lonnet::logthis('correct:'.$submission.':'.$subm{'correct'}); - $submission_data{'_correct'}++; - $submission_data{'_count'}++; + $stats{'correct_count'}++; $submission_data{$subm{'correct'}}->{$submission}->[0]++; } elsif (&numerical_submission_is_incorrect($subm{'award'})) { - &Apache::lonnet::logthis('incorrect:'.$submission.':'.$subm{'correct'}); - $submission_data{'_count'}++; + $stats{'incorrect_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; + $stats{'correct_count'} |= 0; + $stats{'incorrect_count'} |= 0; + $stats{'students'}=scalar(keys(%students)); + return (\%submission_data,\%stats); } sub numerical_submission_is_correct { my ($award) = @_; - &Apache::lonnet::logthis('award = "'.$award.'"'); if ($award =~ /^(APPROX_ANS|EXACT_ANS)$/) { return 1; } else { @@ -508,14 +568,11 @@ sub numerical_bin_table { ''.&mt('Correct').''. ''.&mt('Count').''. ''.$/; - for (my $i=0;$i[$i-1]; - } - my $highnum = $bins->[$i]; + for (my $i=0;$i[$i]; + if ($i == 0) { $lownum = '-∞'; } + my $highnum = $bins->[$i+1]; + if ($i == scalar(@{$bins})-2) { $highnum = '∞'; } $table .= ''. ''.$labels->[$i].''. @@ -535,7 +592,10 @@ sub numerical_determine_answers { my ($r,$resource,$partid,$respid,$students)=@_; my $c = $r->connection(); # - # FIX ME: May need progress dialog updates + my %prog_state=&Apache::lonhtmlcommon::Create_PrgWin + ($r,'Student Answer Compilation Status', + 'Student Answer Compilation Progress', scalar(@$students), + 'inline',undef,'Statistics','stats_status'); # # Read in the cache (if it exists) before we start timing things. &Apache::lonstathelpers::ensure_proper_cache($resource->{'symb'}); @@ -553,13 +613,16 @@ sub numerical_determine_answers { $sdom); # make the key my $key = $partid.'.'.$respid; - $correct->{$sname.':'.$sdom}->{'answer'} = - $analysis->{$key.'.answer'}->[0]; - $correct->{$sname.':'.$sdom}->{'unit'} = - $analysis->{$key.'.unit'}->[0]; + foreach my $item ('answer','unit','ans_high','ans_low') { + $correct->{$sname.':'.$sdom}->{$item} = + $analysis->{$key.'.'.$item}->[0]; + } $answers{$analysis->{$key.'.answer'}->[0]}++; + &Apache::lonhtmlcommon::Increment_PrgWin($r,\%prog_state, + &mt('last student')); } &Apache::lonstathelpers::write_analysis_cache(); + &Apache::lonhtmlcommon::Close_PrgWin($r,\%prog_state); return ($correct,\%answers); } @@ -709,9 +772,9 @@ sub radio_response_analysis { $analysis_html .= $table; # Gather student data my $response_data = &Apache::loncoursedata::get_response_data - (\@Apache::lonstatistics::SelectedSections, + ([&Apache::lonstatistics::get_selected_sections()], $Apache::lonstatistics::enrollment_status, - $resource->{'symb'},$respid); + $resource->symb,$respid); my $correct; # either a hash reference or a scalar if ($problem_analysis->{'answercomputed'} || scalar(@$concepts) > 1) { # This takes a while for large classes... @@ -825,16 +888,14 @@ sub radio_response_analysis { $choice_plot ne '' || $stacked_plot ne '') { my $correct = $foil_choice_data->{'_correct'}; - if (! defined($correct) || $correct eq '') { - $correct = 0; - } - my $incorrect = + $correct |= 0; + my $incorrect = $foil_choice_data->{'_count'}-$correct; $analysis_html.= ''. ''. &mt($pre_graph_text, $plot_num,$foil_choice_data->{'_count'}, $correct, - $foil_choice_data->{'_count'}-$correct, + $incorrect, $foil_choice_data->{'_students'}, @extra_data). ''.$/; @@ -888,7 +949,6 @@ sub ensure_start_end_times { sub RR_concept_plot { my ($concepts,$foil_data,$title) = @_; - &Apache::lonnet::logthis('got to RR_concept_plot'); # my %correct_by_concept; my %incorrect_by_concept; @@ -936,7 +996,6 @@ sub RR_concept_plot { sub RR_create_percent_selected_plot { my ($concepts,$foils,$foil_data,$title) = @_; - &Apache::lonnet::logthis('got to RR_create_percent_selected_plot'); # if ($foil_data->{'_count'} == 0) { return ''; }; my %correct_selections; @@ -1008,7 +1067,6 @@ sub RR_create_percent_selected_plot { sub RR_create_stacked_selection_plot { my ($foils,$foil_data,$title,$true_foils)=@_; # - &Apache::lonnet::logthis('got to RR_create_stacked_selection_plot'); my @dataset; # array of array refs - multicolor rows $datasets[row]->[col] my @labels; my $count; @@ -1119,9 +1177,9 @@ sub OptionResponseAnalysis { $problem->{'respid'}); # Note: part data is not needed. my $PerformanceData = &Apache::loncoursedata::get_response_data - (\@Apache::lonstatistics::SelectedSections, + ([&Apache::lonstatistics::get_selected_sections()], $Apache::lonstatistics::enrollment_status, - $resource->{'symb'},$respid); + $resource->symb,$respid); if (! defined($PerformanceData) || ref($PerformanceData) ne 'ARRAY' ) { $r->print('

'. @@ -1255,6 +1313,7 @@ sub OR_tries_analysis { # Create Foil Plots my $data_count = $response_data{'_total'}->[$try]; my $correct = $response_data{'_correct'}->[$try]; + $correct |= 0; my @Datasets; foreach my $option ('_correct',@{$ORdata->{'_Options'}}) { next if (! exists($foil_plot[$try]->{$option})); @@ -1417,6 +1476,7 @@ sub OR_time_analysis { ## my ($processed_time_data,$correct,$data_count,$student_count) = &OR_time_process_data($performance_data,$begin_index,$end_index); + $correct |= 0; ## $table .= ''. &mt('[_1] submissions from [_2] students, [_3] correct, [_4] incorrect', @@ -1760,7 +1820,6 @@ sub CreateInterface { $Str .= ''; $Str .= ''.&mt('Sections').''; $Str .= ''.&mt('Enrollment Status').''; -# $Str .= ''.&mt('Sequences and Folders').''; $Str .= ' '; $Str .= ''."\n"; ## @@ -1773,17 +1832,6 @@ sub CreateInterface { $Str .= &Apache::lonhtmlcommon::StatusOptions(undef,undef,5); $Str .= ''; # -# $Str .= ''; - my $only_seq_with_assessments = sub { - my $s=shift; - if ($s->{'num_assess'} < 1) { - return 0; - } else { - return 1; - } - }; - &Apache::lonstatistics::MapSelect('Maps','multiple,all',5, - $only_seq_with_assessments); ## ## $Str .= '';