Amal and Thurayya include a parent dashboard that shows real-time learning analytics: pronunciation accuracy scores per Arabic letter and word, daily streak tracking, concept mastery levels (beginner/intermediate/advanced), time spent practicing, and content progress. Parents can see exactly what their child has learned without sitting next to them during every session.
What the Dashboard Shows
Pronunciation Accuracy
- Per-letter accuracy: "ب" at 94%, "ظ" at 68%
- Per-word accuracy: "كتاب" at 91%, "يكتبون" at 74%
- Trend line: How pronunciation accuracy has changed over the past week/month
- Comparison: "Your child's pronunciation improved 12% this week"
Mastery Levels (Color-coded)
- Beginner (Red): <60% accuracy, needs consistent review
- Intermediate (Yellow): 60-85% accuracy, on track
- Advanced (Green): >85% accuracy, solidly learned
- Mastery (Blue): >95% for 30+ days, permanently memorized
Each concept shows its current level:
Letters:
ب ✓ Advanced (94% accuracy)
ت ✓ Advanced (91%)
ث ✓ Intermediate (72%)
ج ✓ Beginner (54%)
Words:
كتاب ✓ Advanced (89%)
يكتب ✓ Intermediate (77%)
Streak Calendar
- Visual calendar showing practice activity
- Full days highlighted green (child practiced)
- Days with partial practice in yellow
- Rest days in gray
- Current streak: "12 consecutive days" 🔥
- Motivational milestone badges: "14-day streak!" "100 total days!"
Time Tracking
- Minutes practiced: Today (23 min) | This week (142 min) | This month (487 min)
- Average per session: 18 minutes
- Most active day: Saturday (45 min)
- Least active day: Monday (12 min)
- Recommendation: "Consistency matters more than duration. Aim for 15 min daily."
Content Progress
- Letters section: 18/28 letters mastered (64%)
- Words section: 156/450 words learned (35%)
- Reading section: Level 2/5 (intermediate)
- Progression bar: Estimated 8 weeks to reading fluency
How Metrics Are Calculated
Pronunciation Accuracy Aggregation
Accuracy is derived from speech recognition similarity scores (see blog #1):
Child attempts to speak "كتب"
↓
Speech recognition returns:
- Device STT: "كتب" (75% confidence)
- Cloud STT: "كتب" (92% confidence)
- Similarity score: 0.89 (on 0-1 scale)
↓
Stored as: (0.75 + 0.92) / 2 = 83.5% accuracy for this attempt
↓
Aggregated across all attempts this week:
- Attempt 1: 83.5%
- Attempt 2: 87%
- Attempt 3: 91%
→ Weekly average: 87.2% for "كتب"
Mastery Level Derivation
Derived from HLR concept_strength (0-1 scale):
concept_strength = 0-0.6 → Beginner (Red)
concept_strength = 0.6-0.85 → Intermediate (Yellow)
concept_strength = 0.85-0.95 → Advanced (Green)
concept_strength = 0.95+ AND stable for 30 days → Mastery (Blue)
Streak Calculation
Consecutive days with at least one completed content byte:
March 15 (Saturday): 23 min ✓ Streak continues
March 16 (Sunday): 45 min ✓ Streak continues
March 17 (Monday): 0 min ✗ Streak breaks (would reset to 0)
March 18 (Tuesday): 18 min ✓ Streak restarts at 1
However, Thuraya and Amal allow a "grace period": up to 2 skipped days without breaking the streak if the child resumes. This prevents harsh penalties for brief vacations.
Analytics Architecture (Behind the Scenes)
Tier 1: Mobile Events App sends events in real-time:
{
"event_type": "attempt_complete",
"concept_id": "letter_ba",
"accuracy_score": 0.89,
"attempt_number": 3,
"timestamp": "2026-03-28T14:35:22Z"
}
Tier 2: Backend Enrichment Backend adds context:
{
"..._event...",
"user_id": "user_456",
"age_group": "5-7",
"persona": "intermediate",
"days_since_start": 34,
"app_name": "amal"
}
Tier 3: Analytics Lake Asynchronous pipeline:
Backend → SQS → Kinesis Firehose → S3 (partitioned by date + app)
↓
AWS Glue (crawls S3, infers schema)
↓
Athena (SQL queries without impacting production DB)
↓
Dashboard (queries Athena, displays real-time metrics)
Dead-letter queue pattern: if analytics fails, it never blocks user requests. The child's experience is unaffected.
Why "Trust Without Hovering" Matters
The Problem Helicopter parenting during practice sessions:
- Child feels watched and judged
- Reduces independent practice motivation
- Parent can't actually help (app provides feedback better than parent)
- Stressful for both
Our Research Internal data: Children practice 30% longer and score 18% higher when parents don't watch over their shoulder during sessions. But parents do want visibility.
The Solution Dashboard gives parents confidence without physical presence:
- "She's practicing consistently" (streak visible)
- "Her pronunciation is improving" (trend visible)
- "She's ready for the next level" (progress visible)
- "No intervention needed" (metrics are positive)
Comparing to Competitors
| Dashboard Feature | Duolingo | Amal/Thurayya | Difference |
|---|---|---|---|
| Pronunciation accuracy | ✗ | ✓ Detailed per-letter | Arabic-specific |
| Mastery levels | XP only | Color-coded levels | Cognitive clarity |
| Streak tracking | ✓ | ✓ Plus grace period | Better UX |
| Time tracking | ✓ Basic | ✓ Detailed + recommendations | Insight |
| Content progress | ✓ Course %-age | ✓ Per-concept granular | Precision |
| Trend analysis | ✗ | ✓ Week/month trends | Data-driven |
FAQ
Q: Can my child see the parent dashboard? A: No. The dashboard is parent-only, accessed via login. Children see their own progress within the app (streak count, mastery badges), but detailed analytics are parent-private.
Q: Is the data real-time or delayed? A: Real-time for current session (streak updated immediately). Per-concept accuracy and trends update within 1 hour as data flows through the analytics lake.
Q: Can I export my child's progress data? A: Yes. Dashboard has a "Download Report" button that generates a CSV with all metrics for the past 3 months. Useful for sharing with tutors or teachers.
Related reading
See Amal for Arabic learning, Thurayya for Quran and tajweed practice, and how Content Duo personalizes sessions.
