Unit Filtering

Effective filtering is essential for focused analysis in InsightLens. This guide shows you how to use filtering tools to drill down into specific units, timeframes, and data segments for meaningful insights.

Overview of Filtering

Purpose of Filtering

  • Focus analysis on specific subsets of data
  • Compare similar units or groups
  • Isolate particular time periods or conditions
  • Reduce complexity in large datasets

Types of Filters Available

  • Unit Characteristics (code, name, discipline)
  • Temporal Filters (year, semester, date ranges)
  • Performance Filters (response rates, satisfaction scores)
  • Geographic Filters (campus, location)
  • Data Quality Filters (completeness, reliability)

Filter Locations

Global Filters (Dashboard)

Available across all pages for consistent analysis:

Time Range Selector

  • Quick options: Last semester, Last year, Custom range
  • Academic year vs Calendar year modes
  • Semester-specific selections
  • Rolling periods (last 6 months, etc.)

Unit Type Filters

  • Undergraduate vs Postgraduate
  • Core vs Elective units
  • Online vs Face-to-face delivery
  • Credit point ranges

Page-Specific Filters

Units Page Filters

  • Advanced search and filtering interface
  • Multiple criteria combination
  • Saved filter presets
  • Bulk action capabilities

Dashboard Widget Filters

  • Individual chart filtering options
  • Quick filter buttons on overview cards
  • Contextual filters based on current view
  • Temporary filters that don’t affect other pages

Unit Characteristic Filters

Unit Code Filtering

Exact Match

  • Enter specific unit code (e.g., “ISYS2001”)
  • Useful for single unit deep-dive analysis
  • Case-insensitive matching
  • Partial code matching available

Pattern Matching

  • Wildcard searches (e.g., “ISYS*” for all ISYS units)
  • Multiple unit codes separated by commas
  • Regular expression support for advanced users
  • Unit family grouping (all 1000-level units)

Discipline and School Filtering

Hierarchical Structure

University
├── Business School
│   ├── Accounting
│   ├── Finance
│   └── Information Systems
├── Engineering
│   ├── Civil Engineering
│   ├── Mechanical Engineering
│   └── Software Engineering
└── Arts and Sciences
    ├── Psychology
    ├── Literature
    └── Mathematics

Multi-Level Selection

  • Choose entire schools or specific disciplines
  • Mix and match different levels
  • Exclude specific subdisciplines
  • Compare across school boundaries

Academic Level Filtering

Standard Levels

  • Foundation (preparatory courses)
  • Undergraduate (1000-3000 level)
  • Postgraduate Coursework (4000-5000 level)
  • Postgraduate Research (6000+ level)

Custom Level Definitions

  • Institution-specific level codes
  • Cross-level analysis capabilities
  • Progression tracking through levels
  • Prerequisite relationship filtering

Temporal Filtering

Academic Calendar Alignment

Semester-Based Filtering

  • Standard semester 1, 2, summer
  • Trimester systems (1, 2, 3)
  • Quarter systems (fall, winter, spring, summer)
  • Custom academic periods

Year Selection Options

  • Single year analysis
  • Multi-year comparisons
  • Academic year vs calendar year
  • Rolling year calculations

Advanced Time Filtering

Relative Time Periods

  • “Last 3 semesters”
  • “Same semester last year”
  • “Previous academic year”
  • “Rolling 12 months”

Specific Date Ranges

  • Custom start and end dates
  • Survey completion date filtering
  • Import date filtering
  • Academic period boundaries

Trend Analysis Time Windows

Comparison Periods

  • Before/after intervention analysis
  • Pre/post curriculum change comparison
  • Instructor change impact assessment
  • Seasonal variation analysis

Performance-Based Filtering

Satisfaction Score Filters

Score Ranges

  • High performers (4.0+ average)
  • Concerning performance (below 3.0)
  • Improvement targets (3.0-3.5)
  • Excellence recognition (4.5+)

Percentile-Based Filtering

  • Top 10% performers
  • Bottom quartile for attention
  • Middle 50% for baseline
  • Outlier identification (beyond 2 standard deviations)

Response Rate Filtering

Rate Thresholds

  • Minimum response rates for reliability
  • High engagement units (>70% response)
  • Low engagement concerns (<30% response)
  • Target achievement (institutional benchmarks)

Statistical Confidence

  • Minimum sample sizes for validity
  • Confidence interval requirements
  • Margin of error considerations
  • Power analysis inclusion criteria

Trend-Based Filtering

Improvement Patterns

  • Units showing consistent improvement
  • Declining performance indicators
  • Stable but concerning patterns
  • Volatile or inconsistent results

Change Magnitude

  • Significant positive changes
  • Concerning negative shifts
  • Minimal change patterns
  • Dramatic transformation cases

Geographic and Delivery Filters

Campus Location

Physical Locations

  • Main campus units
  • Regional campus offerings
  • International campus data
  • Distance education providers

Location-Specific Analysis

  • Resource availability impacts
  • Student demographic differences
  • Local cultural factors
  • Infrastructure variations

Delivery Mode Filtering

Teaching Methods

  • Face-to-face instruction
  • Online delivery
  • Blended/hybrid models
  • Block/intensive modes

Technology Integration

  • High-tech delivery methods
  • Traditional approaches
  • Simulation and lab-based units
  • Field experience components

Data Quality Filters

Completeness Filters

Response Completeness

  • Units with complete survey data
  • Partial response filtering
  • Missing data tolerance levels
  • Data quality scoring

Question Coverage

  • All required questions answered
  • Optional question participation
  • Comment section completion
  • Demographic data availability

Reliability Filters

Sample Size Adequacy

  • Minimum student enrollment
  • Sufficient response numbers
  • Statistical power requirements
  • Confidence level maintenance

Data Consistency

  • Internal consistency checks
  • Cross-validation results
  • Outlier detection and handling
  • Quality assurance passed

Advanced Filtering Techniques

Combined Filter Logic

AND Logic (All conditions must be met)

  • Business School AND Undergraduate AND High Performance
  • Creates narrow, focused datasets
  • Useful for specific case studies
  • Ensures all criteria satisfaction

OR Logic (Any condition can be met)

  • Engineering OR Computer Science units
  • Expands dataset inclusively
  • Useful for broad comparisons
  • Captures related categories

NOT Logic (Exclude specific conditions)

  • All units EXCEPT foundation courses
  • Remove outliers or special cases
  • Focus on standard offerings
  • Clean datasets for analysis

Filter Combinations

Nested Filtering

(Business School OR Engineering) 
AND Undergraduate 
AND (Response Rate > 50%) 
AND (Year = 2023 OR Year = 2024)

Sequential Filtering

  1. Start with broad criteria
  2. Progressively narrow focus
  3. Evaluate results at each step
  4. Fine-tune for optimal dataset

Saved Filter Presets

Common Preset Examples

  • “High Priority Units”: Low satisfaction + High enrollment
  • “Success Stories”: High satisfaction + Improving trends
  • “New Offerings”: First or second year of delivery
  • “Core Curriculum”: Required units for degree programs

Custom Preset Creation

  1. Configure desired filter combination
  2. Save with descriptive name
  3. Share with colleagues if appropriate
  4. Update as criteria evolve

Filter Management

Performance Optimization

Large Dataset Handling

  • Progressive filtering for better performance
  • Index-optimized filter combinations
  • Cached results for common filters
  • Background processing for complex queries

Memory Management

  • Filter result caching
  • Automatic cleanup of temporary filters
  • Efficient data loading strategies
  • Resource usage monitoring

User Experience Features

Filter Persistence

  • Maintain filters across sessions
  • Remember last-used configurations
  • Quick restoration of previous settings
  • Cross-page filter consistency

Visual Feedback

  • Active filter indicators
  • Result count updates
  • Performance impact warnings
  • Filter conflict detection

Best Practices

Effective Filter Strategies

Start Broad, Then Narrow

  1. Begin with general time period
  2. Add major category filters
  3. Refine with performance criteria
  4. Fine-tune with specific requirements

Maintain Context

  • Always show applied filters clearly
  • Provide easy filter removal options
  • Include unfiltered comparisons when relevant
  • Document filter rationale in reports

Common Pitfalls to Avoid

Over-Filtering

  • Too restrictive criteria yielding minimal data
  • Loss of statistical power
  • Missing important patterns
  • Reduced generalizability

Filter Bias

  • Unconsciously selecting favorable data
  • Ignoring inconvenient results
  • Cherry-picking for desired outcomes
  • Inadequate representative sampling

Quality Assurance

Filter Validation

  • Check result counts for reasonableness
  • Verify filter logic accuracy
  • Test with known data subsets
  • Cross-validate with alternative approaches

Documentation Standards

  • Record filter criteria used
  • Note any unusual selections
  • Explain rationale for choices
  • Include filter information in reports

Troubleshooting Filters

Common Issues

No Results Returned

  • Criteria too restrictive
  • Date range issues
  • Data availability problems
  • Filter logic errors

Too Many Results

  • Insufficient filtering
  • Broad criteria combinations
  • Missing quality filters
  • Performance impact concerns

Unexpected Results

  • Filter logic conflicts
  • Data interpretation errors
  • Cache consistency issues
  • System synchronization problems

Diagnostic Steps

  1. Review each filter individually
  2. Test simplified filter combinations
  3. Check data availability for time periods
  4. Verify filter logic interpretation
  5. Clear cache and retry if needed
Previous: Exploring Trends Next: Chart Interactions