Spaces:
Sleeping
Sleeping
Create pages/1_Data_Card_and_Data_collection.py
Browse files
pages/pages/1_Data_Card_and_Data_collection.py
ADDED
|
@@ -0,0 +1,115 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import pandas as pd
|
| 3 |
+
import streamlit as st
|
| 4 |
+
from io import StringIO
|
| 5 |
+
|
| 6 |
+
# Persistent file path for the dataset
|
| 7 |
+
DATA_FILE_PATH = "healthcare-dataset-stroke-data.csv"
|
| 8 |
+
|
| 9 |
+
# Page Title
|
| 10 |
+
st.markdown("<h1 style='text-align:center; color:white;'>healthcare-dataset</h1>", unsafe_allow_html=True)
|
| 11 |
+
|
| 12 |
+
# Function to load the dataset from the disk
|
| 13 |
+
def load_dataset():
|
| 14 |
+
if os.path.exists(DATA_FILE_PATH):
|
| 15 |
+
return pd.read_csv(DATA_FILE_PATH)
|
| 16 |
+
else:
|
| 17 |
+
return None
|
| 18 |
+
|
| 19 |
+
# Load or reload the dataset into session state if not already done
|
| 20 |
+
if "dataset" not in st.session_state:
|
| 21 |
+
st.session_state["dataset"] = load_dataset()
|
| 22 |
+
|
| 23 |
+
# File uploader widget to upload a new dataset
|
| 24 |
+
uploaded_file = st.file_uploader("Choose a CSV file", type=["csv"])
|
| 25 |
+
|
| 26 |
+
if uploaded_file is not None:
|
| 27 |
+
# Read the uploaded CSV file into a pandas DataFrame
|
| 28 |
+
df = pd.read_csv(uploaded_file)
|
| 29 |
+
|
| 30 |
+
# Save the dataset permanently to disk
|
| 31 |
+
df.to_csv(DATA_FILE_PATH, index=False)
|
| 32 |
+
|
| 33 |
+
# Update session state
|
| 34 |
+
st.session_state["dataset"] = df
|
| 35 |
+
|
| 36 |
+
# Display success message
|
| 37 |
+
st.success(f"Dataset uploaded and saved permanently as {DATA_FILE_PATH}!")
|
| 38 |
+
|
| 39 |
+
# Access the dataset from session state
|
| 40 |
+
df = st.session_state.get("dataset")
|
| 41 |
+
|
| 42 |
+
if df is not None:
|
| 43 |
+
st.subheader("Dataset Preview:")
|
| 44 |
+
st.write(df, use_container_width=True)
|
| 45 |
+
|
| 46 |
+
st.subheader("Info of the Dataset:")
|
| 47 |
+
buffer = StringIO()
|
| 48 |
+
df.info(buf=buffer)
|
| 49 |
+
st.text(buffer.getvalue())
|
| 50 |
+
|
| 51 |
+
st.subheader("Dataset Shape (Rows, Columns):")
|
| 52 |
+
st.write(df.shape)
|
| 53 |
+
|
| 54 |
+
st.markdown('''**Dataset :**
|
| 55 |
+
| **Feature** | **Description** | **Example** |
|
| 56 |
+
|--------------------------|--------------------------------------------------------------------|------------------------------|
|
| 57 |
+
| **gender** | Represents the gender of the individual. | "Female","Male" |
|
| 58 |
+
| **age** | Represents the age of the individual in years.. | 45 |
|
| 59 |
+
| **hypertension** | Indicates whether the individual has hypertension . | 1 |
|
| 60 |
+
| **heart_disease** | Indicates whether the individual has a history of heart disease. | 0 |
|
| 61 |
+
| **ever_married** | Indicates whether the individual has ever been married. | "Yes" |
|
| 62 |
+
| **work_type** | Gender of the customer (0 - Male, 1 - Female). | "Self-employed |
|
| 63 |
+
| **Residence_type** |Indicates whether the individual resides in an urban or rural area. | "Urban" |
|
| 64 |
+
| **avg_glucose_level** | Represents the average glucose level in the blood (in mg/dL).. | 132.5 |
|
| 65 |
+
| **bmi** | Represents the individual's Body Mass Index (BMI), calculated as. | 27.4 |
|
| 66 |
+
| **smoking_status** | Represents the individual's smoking habits | "Never smoked" |
|
| 67 |
+
| **stroke** | Indicates whether the individual has had a stroke. | 0 (No), 1 (Yes) |
|
| 68 |
+
''')
|
| 69 |
+
|
| 70 |
+
|
| 71 |
+
|
| 72 |
+
else:
|
| 73 |
+
st.info("No dataset found. Please upload a CSV file.")
|
| 74 |
+
|
| 75 |
+
# Define the URL of the background image (use your own image URL)
|
| 76 |
+
background_image_url = "https://cdn-uploads.huggingface.co/production/uploads/67441c51a784a9d15cb12871/xpoN_mbctlrQAgRU06EPt.jpeg"
|
| 77 |
+
|
| 78 |
+
# Apply custom CSS for the background image and overlay
|
| 79 |
+
st.markdown(
|
| 80 |
+
f"""
|
| 81 |
+
<style>
|
| 82 |
+
.stApp {{
|
| 83 |
+
background-image: url("{background_image_url}");
|
| 84 |
+
background-size: cover;
|
| 85 |
+
background-position: center;
|
| 86 |
+
height: 100vh;
|
| 87 |
+
}}
|
| 88 |
+
|
| 89 |
+
/* Semi-transparent overlay */
|
| 90 |
+
.stApp::before {{
|
| 91 |
+
content: "";
|
| 92 |
+
position: absolute;
|
| 93 |
+
top: 0;
|
| 94 |
+
left: 0;
|
| 95 |
+
width: 100%;
|
| 96 |
+
height: 100%;
|
| 97 |
+
background: rgba(0, 0, 0, 0.4);
|
| 98 |
+
z-index: -1;
|
| 99 |
+
}}
|
| 100 |
+
|
| 101 |
+
/* Styling the content to ensure text visibility */
|
| 102 |
+
.stMarkdown {{
|
| 103 |
+
color: white;
|
| 104 |
+
font-size: 100px;
|
| 105 |
+
}}
|
| 106 |
+
</style>
|
| 107 |
+
""",
|
| 108 |
+
unsafe_allow_html=True
|
| 109 |
+
)
|
| 110 |
+
|
| 111 |
+
|
| 112 |
+
if st.button("Previous ⏮️"):
|
| 113 |
+
st.switch_page("pages/0_Problem-Statement_and_Aim.py")
|
| 114 |
+
if st.button("Next ⏭️"):
|
| 115 |
+
st.switch_page("pages/2_Data_CLeaning_and_Preprocessing.py")
|