Understanding Moods

Reflect AI automatically analyzes the emotional tone of your journal entries using VADER (Valence Aware Dictionary and sEntiment Reasoner), a sophisticated sentiment analysis tool designed for social media and personal text.

​

How Sentiment Analysis Works

When you save an entry, Reflect AI:

​

The Five Mood Categories

​

😊 Very Positive (Joyful)

Compound Score: ≥ 0.5 What it means: Your entry contains strong positive sentiment with words like “amazing,” “love,” “thrilled,” or “wonderful.” These are your best days! Example phrases:

• “I’m so grateful for today”
• “Everything went perfectly”
• “Feeling absolutely wonderful”
• “Best day in a long time”

AI Response: “What a wonderful day! Your positivity shines through your words.”

---

​

🙂 Positive (Content)

Compound Score: 0.2 to 0.5 What it means: Your entry leans positive with words like “good,” “nice,” “enjoyed,” or “pleasant.” These are solid, satisfying days. Example phrases:

• “Had a productive day”
• “Enjoyed dinner with friends”
• “Things went well”
• “Feeling pretty good”

AI Response: “It sounds like things are going well. Keep nurturing those good moments.”

---

​

😐 Neutral (Balanced)

Compound Score: -0.2 to 0.2 What it means: Your entry is balanced or factual, without strong emotional language. These entries might be informational or reflective. Example phrases:

• “Went to work, came home”
• “Normal day, nothing special”
• “Worked on project, had lunch”
• “Thinking about next week”

AI Response: “Thank you for reflecting today. Every entry helps you understand yourself better.”

---

​

😔 Negative (Reflective)

Compound Score: -0.5 to -0.2 What it means: Your entry contains some challenging emotions with words like “difficult,” “tired,” “stressed,” or “frustrated.” These are tough but manageable days. Example phrases:

• “Feeling a bit overwhelmed”
• “Difficult meeting today”
• “Tired and stressed”
• “Not my best day”

AI Response: “It seems like today had some challenges. Remember, it’s okay to have difficult days.”

---

​

😢 Very Negative (Processing)

Compound Score: ≤ -0.5 What it means: Your entry reflects significant distress with words like “terrible,” “hopeless,” “devastated,” or “awful.” These are the hardest days. Example phrases:

• “Everything went wrong”
• “Feeling really down”
• “Terrible day, nothing worked”
• “So frustrated and upset”

AI Response: “I hear that today was tough. Writing about it is a brave step. Be gentle with yourself.”

​

Understanding Your Scores

You can see detailed sentiment scores in the Insights tab. Here’s how to interpret them:

​

Factors That Influence Mood Detection

​

What VADER Recognizes

​

What Doesn’t Affect Scores

• Entry length: A short positive entry and long positive entry can have similar scores
• Grammar: Spelling and grammar errors don’t impact sentiment detection
• Topics: Writing about “work” isn’t inherently positive or negative—it depends on the emotional words you use

​

Mood Trends Over Time

Reflect AI tracks your mood patterns to help you understand your emotional journey: In the Insights Tab:

• Mood Trend Chart: See your compound scores over the past month
• Mood Distribution: Pie chart showing your mood category breakdown
• Day of Week Patterns: Discover which days tend to be better or harder

Pattern Insights:

• The app identifies activities that correlate with better moods
• Weekly and monthly summaries highlight emotional trends
• Activity-mood correlations show what helps you feel your best

​

Mood vs. Mental Health

It’s important to understand what mood tracking can and cannot tell you: Mood tracking helps you:

• Notice patterns in your emotional life
• Identify activities that lift your spirits
• Track the effectiveness of lifestyle changes
• Reflect on your emotional growth over time

Mood tracking does NOT:

• Diagnose mental health conditions
• Replace professional assessment
• Predict future mental health episodes
• Provide medical or therapeutic advice

​

Privacy and Your Mood Data

Your mood scores are:

• ✅ Calculated locally on your device
• ✅ Never sent to external servers
• ✅ Stored only in your local journal file
• ✅ Completely private to you
• ✅ Exportable with your journal data

No one else can see your mood trends—not even the Reflect AI developers.

​

Improving Mood Accuracy

For the most accurate sentiment analysis:

1. Write naturally: Don’t try to “game” the system. Authentic writing produces meaningful insights.
2. Use emotional words: “Felt great” is clearer than “It was okay” if you’re trying to capture positive feelings.
3. Include context: “Stressful meeting but proud of how I handled it” captures complexity better than just “Stressful.”
4. Be specific: Details about what made a day good or bad help the analysis work better.

​

Next Steps