Build a Real-Time Sentiment Dashboard with RoBERTa

1. 1
   Install Dependencies

   Set up your Python environment. You'll need `transformers` and `torch` to run the model, `scipy` for post-processing, and `streamlit` to build the interactive web dashboard.
2. 2
   Load the Pre-Trained Model

   Download and load the `twitter-roberta-base-sentiment` model and its corresponding tokenizer from the Hugging Face Hub. This model is specifically fine-tuned for sentiment on short texts like social media posts.
3. 3
   Create a Sentiment Scoring Function

   Define a function that takes raw text, tokenizes it for the model, runs inference, and uses a softmax function to convert the output logits into a human-readable probability distribution for negative, neutral, and positive sentiment.
4. 4
   Build the Streamlit Dashboard UI

   Use Streamlit to create the user interface for your Sentiment Dashboard. Add a title, a text area for user input, and a button. When the button is clicked, it will trigger the sentiment analysis function.

import streamlit as st
import pandas as pd
from transformers import AutoModelForSequenceClassification, AutoTokenizer
from scipy.special import softmax

# --- Model Loading (cached for performance) ---
@st.cache_resource
def load_model():
    MODEL = "cardiffnlp/twitter-roberta-base-sentiment"
    tokenizer = AutoTokenizer.from_pretrained(MODEL)
    model = AutoModelForSequenceClassification.from_pretrained(MODEL)
    return tokenizer, model

tokenizer, model = load_model()

# --- Sentiment Scoring Function ---
def get_sentiment_scores(text):
    """Analyzes text and returns sentiment scores."""
    encoded_input = tokenizer(text, return_tensors='pt')
    output = model(**encoded_input)
    scores = output[0][0].detach().numpy()
    scores = softmax(scores)
    labels = ['negative', 'neutral', 'positive']
    return dict(zip(labels, scores))

# --- Streamlit App UI ---
st.set_page_config(layout="wide")
st.title("Sentiment Dashboard")
st.markdown("Analyze text sentiment in real-time using a fine-tuned RoBERTa model.")

user_input = st.text_area("Enter Text for Analysis (e.g., a tweet, review, or comment):", "Streamlit makes building data apps so easy!", height=150)

if st.button("Analyze Sentiment"):
    if user_input:
        with st.spinner('Analyzing...'):
            scores = get_sentiment_scores(user_input)
            
            # Find the dominant sentiment
            dominant_sentiment = max(scores, key=scores.get)
            dominant_score = scores[dominant_sentiment]

            st.subheader("Analysis Results")
            col1, col2 = st.columns(2)

            with col1:
                st.metric(
                    label="Dominant Sentiment", 
                    value=dominant_sentiment.capitalize(),
                    help=f"The highest score was {dominant_score:.2%}"
                )
                st.write("**Full Score Distribution:**")
                st.write(scores)

            with col2:
                st.write("**Score Visualization:**")
                chart_data = pd.DataFrame([scores])
                st.bar_chart(chart_data)
    else:
        st.warning("Please enter some text to analyze.")

# To run: 
# 1. Save this code as `app.py`
# 2. Run `pip install streamlit pandas transformers torch scipy`
# 3. Run `streamlit run app.py` in your terminal