A sentiment analysis dashboard that classifies customer reviews as positive, negative, or neutral using a free HuggingFace model, with a Streamlit UI to upload and analyse in bulk.
A pre-trained DistilBERT model classifies text as POSITIVE or NEGATIVE with a confidence score.
from transformers import pipeline
# Downloads ~700 MB on first run, then cached locally
sentiment = pipeline(
'sentiment-analysis',
model='distilbert-base-uncased-finetuned-sst-2-english',
truncation=True,
max_length=512,
)
reviews = [
"The product quality exceeded my expectations. Fast delivery too!",
"Terrible customer service. Waited 3 weeks and still no refund.",
"It is okay. Nothing special but does the job.",
"Absolutely love this! Best purchase I have made this year.",
"Would not recommend. Broke after two weeks of light use.",
]
results = sentiment(reviews)
for review, r in zip(reviews, results):
print(f"[{r['label']:8} {r['score']:.2%}] {review[:60]}")[POSITIVE 99.82%] The product quality exceeded my expectations. Fast deli [NEGATIVE 99.97%] Terrible customer service. Waited 3 weeks and still no [POSITIVE 56.34%] It is okay. Nothing special but does the job. [POSITIVE 99.92%] Absolutely love this! Best purchase I have made this ye [NEGATIVE 99.86%] Would not recommend. Broke after two weeks of light use.
A simple UI to upload a CSV of reviews and see the sentiment breakdown.
# app.py
import streamlit as st
import pandas as pd
import matplotlib.pyplot as plt
from transformers import pipeline
@st.cache_resource
def load_model():
return pipeline(
'sentiment-analysis',
model='distilbert-base-uncased-finetuned-sst-2-english',
truncation=True, max_length=512,
)
st.title("Sentiment Analysis Dashboard")
st.write("Upload a CSV with a `review` column to analyse customer sentiment.")
uploaded = st.file_uploader("Choose a CSV file", type='csv')
if uploaded:
df = pd.read_csv(uploaded)
if 'review' not in df.columns:
st.error("CSV must have a 'review' column"); st.stop()
st.write(f"Analysing {len(df)} reviews...")
sentiment = load_model()
with st.spinner("Running sentiment analysis..."):
results = sentiment(df['review'].tolist())
df['Sentiment'] = [r['label'] for r in results]
df['Score'] = [round(r['score'], 3) for r in results]
# Summary counts
counts = df['Sentiment'].value_counts()
col1, col2 = st.columns(2)
col1.metric("Positive", counts.get('POSITIVE',0))
col2.metric("Negative", counts.get('NEGATIVE',0))
# Pie chart
fig, ax = plt.subplots()
ax.pie(counts, labels=counts.index, autopct='%1.0f%%',
colors=['#22c55e','#ef4444'])
st.pyplot(fig)
# Full table
st.dataframe(df[['review','Sentiment','Score']])
st.download_button("Download Results", df.to_csv(index=False), "results.csv")# Run with:
streamlit run app.pyA functional analytics tool ready for a real business use case. Upload a month of product reviews, see the trend, identify the most negative feedback, and filter for low-confidence predictions that need human review.
Spotted a bug, broken code, or something that doesn't look right? Tell us what's off and we'll fix it.