Add homepage stats widget with analytics dashboard and mock data support

[arnavlohiya]

Summary

Implements homepage statistics display showing conversation count, memory count, and word count with a detailed analytics dashboard.

Closes #3469

Features Implemented

Stats Banner (Homepage)

• Displays conversations count, memories count, and words count
• Card-style design with rounded corners and chevron indicator
• Tappable banner opens detailed analytics modal
• Clean, minimal design that fits homepage layout

Analytics Dashboard

• TikTok-style vertical swipe interface with card navigation
• Weekly trends: Week-over-week comparison with percentage changes
• Daily cadence: Bar chart showing conversations per day
• Word cloud: Most-used words with 145+ stopwords filtered
• Category breakdown: Pie chart showing conversation categories
• Time-of-day analysis: When users talk most (Morning/Afternoon/Evening/Night)
• Streak tracking: Consecutive days and best day identification
• Efficiency metrics: WPM, memories per conversation, avg words

Mock Data Support

Added DevConstants.useMockData toggle in lib/utils/constants.dart for simulator/emulator testing:

• ✅ Enables UI testing without microphone/backend access
• ✅ Includes 7 realistic mock conversations (~120 words)
• ✅ Includes 15 mock memories across categories
• ✅ Helps contributors test features without physical devices
• ❌ Note: AI Chat won't have memory context with mock data (backend queries its own database)

BEHAVIOR SUMMARY:

• useMockData=true + NO device connected = Mock data shown ✅
• useMockData=true + Device connected = Real device data shown ✅
• useMockData=false = Always uses real API/device data ✅

Files Changed

• lib/pages/home/widgets/stats_widget.dart - Stats banner component with card UI and chevron
• lib/pages/home/widgets/stats_detail_sheet.dart - Analytics dashboard modal (938 lines)
• lib/utils/constants.dart - Mock data configuration toggle
• lib/pages/home/page.dart - Integrated stats widget
• lib/providers/conversation_provider.dart - Added mock conversation data
• lib/providers/memories_provider.dart - Added mock memories data

Screenshots

Testing

• Stats display correctly on homepage
• All three metrics (conversations, memories, words) calculate properly
• Card UI with chevron indicates interactivity
• Modal opens smoothly on banner tap
• Vertical swipe navigation works between analytics cards
• Mock data loads when useMockData = true
• Real data works when useMockData = false
• Word cloud filters stopwords correctly (145+ words)
• Charts render with accurate data
• No compilation errors

Impact

This feature provides users with visible progress metrics to "feel the value from the app everytime they open it" (as requested in #3469), encouraging engagement and sharing.

[gemini-code-assist]

Code Review

This pull request introduces a valuable homepage statistics widget and a detailed analytics dashboard, significantly enhancing user engagement. The implementation of mock data for testing is a commendable practice. My review focuses on improving performance and type safety in the new analytics features. I've identified some critical and high-severity issues where expensive calculations are performed inefficiently and unsafe type casting could lead to runtime errors. Addressing these by moving logic to providers and utilizing data classes will make the new features more robust and performant.

[gemini-code-assist]

Using Map<String, dynamic> and then casting with as is not type-safe and can lead to runtime crashes if a key is misspelled, a value is missing, or the type is incorrect. This makes the code brittle and hard to maintain.

A much safer and more robust approach is to define a dedicated data class to hold the statistics. This provides compile-time type safety, enables autocompletion in the IDE, and makes the code easier to read and reason about.

Here's an example of what the data class could look like:

class _AnalyticsStats {
  final int thisWeekConvos;
  final int lastWeekConvos;
  final int thisWeekWords;
  final int lastWeekWords;
  final Map<String, int> dailyCounts;
  final List<MapEntry<String, int>> topCategories;
  final int streak;
  final MapEntry<String, int>? bestDay;
  final List<MapEntry<String, int>> topWords;
  final Map<String, int> timeOfDayBuckets;
  final double avgWordsPerConvo;
  final double avgWpm;
  final double longestDurationMinutes;
  final int maxWordsInConvo;
  final double memoriesPerConvo;
  final List<_CategoryMomentum> categoryMomentum;

  _AnalyticsStats({
    required this.thisWeekConvos,
    required this.lastWeekConvos,
    required this.thisWeekWords,
    required this.lastWeekWords,
    required this.dailyCounts,
    required this.topCategories,
    required this.streak,
    this.bestDay,
    required this.topWords,
    required this.timeOfDayBuckets,
    required this.avgWordsPerConvo,
    required this.avgWpm,
    required this.longestDurationMinutes,
    required this.maxWordsInConvo,
    required this.memoriesPerConvo,
    required this.categoryMomentum,
  });
}

The _buildStats method should be updated to return an instance of this class, and the build method can then access the properties safely (e.g., stats.dailyCounts).

[gemini-code-assist]

The _buildStats() method is called directly within the build() method. This is inefficient as _buildStats() performs expensive computations over all conversations, and the build() method can be called multiple times by the framework, leading to unnecessary recalculations. To improve performance, these stats should be computed once when the widget is initialized and stored in the state.

I recommend calculating the stats in initState() and storing them in a late final variable. This ensures the heavy computation runs only once when the sheet is first built.

[arnavlohiya]

Solved in the new Commit

[gemini-code-assist]

The _calculateTotalWords method is executed within the build method, which is inefficient as it iterates over all conversations on every rebuild. This can lead to performance degradation, especially with a large number of conversations. This logic should be moved to the ConversationProvider to be calculated once and cached, ensuring the UI remains responsive.

You should remove the _calculateTotalWords method from this widget and add a totalWordsCount getter to ConversationProvider.

// In ConversationProvider:
int get totalWordsCount {
  return conversations.fold(0, (sum, convo) {
    return sum + convo.transcriptSegments.fold(0, (sum, seg) {
      final trimmed = seg.text.trim();
      return sum + (trimmed.isEmpty ? 0 : trimmed.split(RegExp(r'\s+')).length);
    });
  });
}

Suggested change

	final wordsCount = _calculateTotalWords(conversationProvider.conversations);
	final wordsCount = conversationProvider.totalWordsCount;

References

1. When separating UI and provider logic, provider methods should handle state management, while the UI layer is responsible for invoking subsequent asynchronous actions after the state has been modified.

[arnavlohiya]

This update is required as the numbers need to change when conversations are updated in the data. Otherwise, the numbers would reflect old data while conversations are updated.

[beastoin]

wow btw keep it as draft until it's ready

you should seeking for the direct feedback from Aarav man (whom write the ticket)

thank you @arnavlohiya

[mdmohsin7]

closing it for now since there has been no update for 2 weeks, pls feel free to reopen it once its ready

[github-actions]

Hey @arnavlohiya 👋

Thank you so much for taking the time to contribute to Omi! We truly appreciate you putting in the effort to submit this pull request.

After careful review, we've decided not to merge this particular PR. Please don't take this personally — we genuinely try to merge as many contributions as possible, but sometimes we have to make tough calls based on:

• Project standards — Ensuring consistency across the codebase
• User needs — Making sure changes align with what our users need
• Code best practices — Maintaining code quality and maintainability
• Project direction — Keeping aligned with our roadmap and vision

Your contribution is still valuable to us, and we'd love to see you contribute again in the future! If you'd like feedback on how to improve this PR or want to discuss alternative approaches, please don't hesitate to reach out.

Thank you for being part of the Omi community! 💜