app.py

import streamlit as st
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.linear_model import LinearRegression
import re

# Load the CSV file
data = pd.read_csv('amazon.csv', delimiter=',')

# Function to extract numeric values from a string (remove non-numeric characters)
def extract_numeric_value(value):
    numeric_value = re.sub(r'[^\d.]', '', value)
    return float(numeric_value) if numeric_value else None

# Convert 'actual_price' and 'rating' columns to numeric by removing non-numeric characters
data['actual_price'] = data['actual_price'].apply(extract_numeric_value)
data['rating'] = data['rating'].apply(extract_numeric_value)

# Title
st.title('Sales Forecasting and Product Analysis')

# Sidebar filters
category_filter = st.sidebar.selectbox('Select Category', data['category'].unique())
product_filter = st.sidebar.selectbox('Select Product', data[data['category'] == category_filter]['product_name'].unique())

filtered_data = data[(data['category'] == category_filter) & (data['product_name'] == product_filter)]

# Display basic product information
st.subheader('Product Information')
st.write('Product ID:', filtered_data['product_id'].values[0])
st.write('Product Name:', product_filter)
st.write('Category:', category_filter)
st.write('Discounted Price:', filtered_data['discounted_price'].values[0])
st.write('Actual Price:', filtered_data['actual_price'].values[0])
st.write('Discount Percentage:', filtered_data['discount_percentage'].values[0])
st.write('Average Rating:', filtered_data['rating'].mean())
st.write('Rating Count:', filtered_data['rating_count'].values[0])
st.write('About Product:', filtered_data['about_product'].values[0])


# Display product image
product_image_url = filtered_data['img_link'].values[0]
st.image(product_image_url, caption='Product Image', use_column_width=True)


# Cumulative sales plot over rows (assumes rows are ordered chronologically)
cumulative_sales_over_rows = filtered_data.index + 1
st.subheader('Cumulative Sales Over Rows')
fig, ax = plt.subplots(figsize=(10, 6))
ax.plot(cumulative_sales_over_rows, filtered_data['actual_price'].cumsum())
ax.set_xlabel('Rows')
ax.set_ylabel('Cumulative Sales')
st.pyplot(fig)

# Sales forecasting using Linear Regression
st.subheader('Sales Forecasting')
forecast_days = st.number_input('Enter the number of days for sales forecasting:', min_value=1, value=7)

# Convert the row index to numeric values
row_numeric = cumulative_sales_over_rows.values.reshape(-1, 1)
X = row_numeric
y = filtered_data['actual_price'].cumsum().values
model = LinearRegression()
model.fit(X, y)

# Generate future row indices for forecasting
last_row = cumulative_sales_over_rows.max()
forecast_rows = [last_row + i for i in range(1, forecast_days + 1)]
forecast_rows_numeric = pd.Series(forecast_rows).values.reshape(-1, 1)

# Predict cumulative sales for forecasted rows
forecast_values = model.predict(forecast_rows_numeric)
forecast_cumulative_sales = forecast_values

forecast_df = pd.DataFrame({'Row': forecast_rows, 'Forecasted Cumulative Sales': forecast_cumulative_sales})
st.write(forecast_df)


# Footer
st.text('Data Used: Amazon Sales Data')