app.py

import streamlit as st
import requests
import os
import json
import pandas as pd
import plotly.graph_objects as go
import plotly.express as px
import time
from datetime import datetime, timedelta
import random
import nltk
from nltk.corpus import stopwords
from wordcloud import WordCloud
import matplotlib.pyplot as plt
from streamlit_lottie import st_lottie
import base64
from PIL import Image
import io

# Download NLTK data
nltk.download('stopwords', quiet=True)
nltk.download('punkt', quiet=True)

# Set page config
st.set_page_config(page_title="DSEAP", page_icon="📊", layout="wide")

# Custom CSS for styling
st.markdown("""
    <style>
        @import url('https://fonts.googleapis.com/css2?family=Roboto:wght@300;400;700&display=swap');
        
        .stApp {
            background: #f0f2f6;
            font-family: 'Roboto', sans-serif;
        }
        .header {
            font-size: 42px;
            font-weight: bold;
            color: #1e3799;
            text-align: center;
            margin-bottom: 30px;
            text-shadow: 2px 2px 4px rgba(0,0,0,0.1);
        }
        .subheader {
            font-size: 28px;
            font-weight: bold;
            color: #3c6382;
            text-align: center;
            margin-bottom: 25px;
        }
        .section {
            background: white;
            padding: 25px;
            border-radius: 15px;
            box-shadow: 0px 8px 16px rgba(0, 0, 0, 0.1);
            margin-bottom: 30px;
        }
        .footer {
            text-align: center;
            font-size: 14px;
            color: #555;
            margin-top: 30px;
            padding: 10px;
            border-top: 1px solid #ddd;
        }
        .stProgress > div > div > div > div {
            background-image: linear-gradient(to right, #4a69bd, #6a89cc);
        }
        .stButton>button {
            background-color: #4a69bd;
            color: white;
            font-weight: bold;
            border-radius: 5px;
            padding: 10px 20px;
            transition: all 0.3s ease;
        }
        .stButton>button:hover {
            background-color: #1e3799;
            box-shadow: 0px 4px 8px rgba(0, 0, 0, 0.2);
        }
        .animated-div {
            background: linear-gradient(-45deg, #4a69bd, #6a89cc, #54a0ff, #5f27cd);
            background-size: 400% 400%;
            animation: gradient 15s ease infinite;
            padding: 15px;
            border-radius: 10px;
            margin-bottom: 15px;
            color: white;
            font-weight: bold;
        }
        @keyframes gradient {
            0% {background-position: 0% 50%;}
            50% {background-position: 100% 50%;}
            100% {background-position: 0% 50%;}
        }
        .info-box {
            background-color: #e3f2fd;
            border-left: 5px solid #2196f3;
            padding: 15px;
            margin-bottom: 20px;
            border-radius: 5px;
        }
        .warning-box {
            background-color: #fff3e0;
            border-left: 5px solid #ff9800;
            padding: 15px;
            margin-bottom: 20px;
            border-radius: 5px;
        }
    </style>
""", unsafe_allow_html=True)

# Function to load Lottie animations
def load_lottieurl(url: str):
    r = requests.get(url)
    if r.status_code != 200:
        return None
    return r.json()

# Load Lottie animations
lottie_analytics = load_lottieurl("https://assets5.lottiefiles.com/packages/lf20_qp1q7mct.json")
lottie_skills = load_lottieurl("https://assets5.lottiefiles.com/private_files/lf30_wqypnpu5.json")

# Function to call the Together AI model
def call_ai_model(prompt):
    url = "https://api.together.xyz/v1/chat/completions"
    payload = {
        "model": "NousResearch/Nous-Hermes-2-Yi-34B",
        "temperature": 1.05,
        "top_p": 0.9,
        "top_k": 50,
        "repetition_penalty": 1,
        "n": 1,
        "messages": [{"role": "user", "content": prompt}],
        "stream_tokens": True,
    }

    TOGETHER_API_KEY = os.getenv('TOGETHER_API_KEY')
    if TOGETHER_API_KEY is None:
        raise ValueError("TOGETHER_API_KEY environment variable not set.")
    
    headers = {
        "accept": "application/json",
        "content-type": "application/json",
        "Authorization": f"Bearer {TOGETHER_API_KEY}",
    }

    try:
        response = requests.post(url, json=payload, headers=headers, stream=True)
        response.raise_for_status()
        
        full_response = ""
        for line in response.iter_lines():
            if line:
                line_content = line.decode('utf-8')
                if line_content.startswith("data: "):
                    line_content = line_content[6:]
                try:
                    json_data = json.loads(line_content)
                    if "choices" in json_data:
                        delta = json_data["choices"][0]["delta"]
                        if "content" in delta:
                            full_response += delta["content"]
                except json.JSONDecodeError:
                    continue
        
        return full_response.strip()
    except requests.exceptions.RequestException as e:
        st.error(f"API call failed: {e}")
        return "An error occurred while fetching AI insights."

# Function to generate simulated data
def generate_simulated_data(categories, count=5):
    return {cat: [random.randint(1, 100) for _ in range(count)] for cat in categories}

# Function to create word cloud
def create_word_cloud(text):
    stopwords = set(stopwords.words('english'))
    wordcloud = WordCloud(width=800, height=400, background_color='white', stopwords=stopwords).generate(text)
    
    fig, ax = plt.subplots(figsize=(10, 5))
    ax.imshow(wordcloud, interpolation='bilinear')
    ax.axis('off')
    return fig

# Function to create a downloadable Excel report
def create_excel_report(data):
    output = io.BytesIO()
    with pd.ExcelWriter(output, engine='xlsxwriter') as writer:
        for sheet_name, df in data.items():
            df.to_excel(writer, sheet_name=sheet_name, index=False)
    output.seek(0)
    return output

# Function for animated progress bar
def animated_progress_bar():
    progress_bar = st.progress(0)
    status_text = st.empty()
    phrases = [
        "Empowering youth with digital skills...",
        "Unlocking your potential...",
        "Building a brighter future...",
        "Connecting skills to opportunities...",
        "Transforming lives through technology..."
    ]
    for i in range(100):
        progress_bar.progress(i + 1)
        status_text.text(random.choice(phrases))
        time.sleep(0.05)
    status_text.text("Ready to empower!")
    time.sleep(0.5)
    status_text.empty()
    progress_bar.empty()

# Streamlit app layout
st.markdown('<div class="header">Digital Skills and Employment Analytics Platform (DSEAP)</div>', unsafe_allow_html=True)
st.markdown('<div class="subheader">Empowering Youth Through Data-Driven Insights</div>', unsafe_allow_html=True)

# Sidebar for navigation
st.sidebar.title("Navigation")
page = st.sidebar.radio("Go to", ["Home", "Skills Analysis", "Program Evaluation", "Barrier Identification", "Recommendations", "Reports"])

if page == "Home":
    animated_progress_bar()
    col1, col2 = st.columns([2, 1])
    with col1:
        st.markdown('<div class="section">', unsafe_allow_html=True)
        st.subheader("Welcome to DSEAP")
        st.write("""
        The Digital Skills and Employment Analytics Platform (DSEAP) is an AI-driven tool designed to enhance 
        the evaluation and improvement of digital empowerment programs in Kenya. Our platform provides comprehensive 
        insights into the demand for digital skills, the effectiveness of existing programs, and the barriers faced 
        by youth in accessing digital opportunities. **A study In cmputing and prgramming**
        """)
        st.markdown('</div>', unsafe_allow_html=True)
    with col2:
        st_lottie(lottie_analytics, height=300, key="analytics")

    st.markdown('<div class="section">', unsafe_allow_html=True)
    st.subheader("Key Features")
    col1, col2, col3 = st.columns(3)
    with col1:
        st.markdown("### 📊 Skills Demand Analysis")
        st.write("Analyze job market trends and identify in-demand digital skills.")
    with col2:
        st.markdown("### 📈 Program Effectiveness")
        st.write("Evaluate the impact of digital empowerment programs on youth employment.")
    with col3:
        st.markdown("### 🚧 Barrier Identification")
        st.write("Identify and address obstacles to digital skill acquisition and employment.")
    st.markdown('</div>', unsafe_allow_html=True)

    st.markdown('<div class="info-box">', unsafe_allow_html=True)
    st.info("📌 **Tip:** Use the sidebar to navigate through different sections of the platform.")
    st.markdown('</div>', unsafe_allow_html=True)

elif page == "Skills Analysis":
    animated_progress_bar()
    st.markdown('<div class="section">', unsafe_allow_html=True)
    st.subheader("Digital Skills Demand Analysis")
    
    col1, col2 = st.columns([3, 1])
    with col1:
        skills = st.multiselect("Select skills to analyze:", 
                                ["Data Analysis", "Web Development", "Digital Marketing", "Cybersecurity", "Cloud Computing", "AI/Machine Learning", "Mobile App Development", "UI/UX Design"])
        regions = st.multiselect("Select regions:", 
                                 ["Kisii", "Nyamira", "Nairobi", "Mombasa", "Kisumu", "Nakuru", "Eldoret", "Thika", "Malindi", "Kitale"])
        time_period = st.selectbox("Select time period:", ["Last 3 months", "Last 6 months", "Last year"])
    with col2:
        st_lottie(lottie_skills, height=200, key="skills")
    
    if st.button("Analyze Skills Demand"):
        with st.spinner("Analyzing skills demand..."):
            # Simulated data generation (optimized)
            demand_data = {skill: [random.randint(1, 100) for _ in regions] for skill in skills}
            
            # Plotting (single chart instead of multiple)
            df = pd.DataFrame(demand_data, index=regions)
            fig = px.bar(df, barmode='group', title="Digital Skills Demand by Region")
            st.plotly_chart(fig, use_container_width=True)
            
            # AI Insights (with timeout)
            ai_prompt = f"Briefly analyze the demand for {', '.join(skills)} in {', '.join(regions)} over the {time_period}. You can be negative on some skills too, just be free. Provide key insights on trends and gaps. Each region with its analysis and use percentages sometimes"
            try:
                with st.spinner("Generating AI insights..."):
                    ai_insights = call_ai_model(ai_prompt)
                st.markdown("### AI Insights")
                st.write(ai_insights)
            except Exception as e:
                st.error(f"Failed to generate AI insights: {str(e)}")
            
            # Word Cloud (optional)
            if st.checkbox("Generate Word Cloud"):
                word_cloud_text = " ".join(skills + regions + ai_insights.split())
                st.markdown("### Skills Demand Word Cloud")
                st.pyplot(create_word_cloud(word_cloud_text))
            
            # Skills growth projection (simplified)
            st.markdown("### Skills Growth Projection")
            growth_data = {skill: [random.uniform(0, 15) for _ in range(2)] for skill in skills}
            years = [datetime.now().year, datetime.now().year + 5]
            df_growth = pd.DataFrame(growth_data, index=years)
            fig_growth = px.line(df_growth, title="Projected Skills Growth (5 Year Projection)")
            st.plotly_chart(fig_growth, use_container_width=True)
    
    st.markdown('</div>', unsafe_allow_html=True)
            


elif page == "Program Evaluation":
    animated_progress_bar()
    st.markdown('<div class="section">', unsafe_allow_html=True)
    st.subheader("Digital Empowerment Program Evaluation")
    
    programs = st.multiselect("Select programs to evaluate:", 
                              ["Ajira Digital", "Remotasks", "Upwork", "Fiverr", "Andela", "Emobiles", "snaphub", "iBizAfrica"])
    metrics = st.multiselect("Select evaluation metrics:", 
                             ["Employment Rate", "Income Increase", "Skill Proficiency", "Program Completion Rate", "Job Satisfaction", "Career Advancement"])
    
    if st.button("Evaluate Programs"):
        with st.spinner("Evaluating programs..."):
            # Simulated data generation
            evaluation_data = generate_simulated_data(programs)
            
            # Plotting
            df = pd.DataFrame(evaluation_data)
            fig = px.line(df, x=df.index, y=df.columns, labels={'index': 'Time (months)', 'value': 'Performance'}, title="Program Performance Over Time")
            fig.update_layout(legend_title_text='Programs')
            st.plotly_chart(fig)
            
            # AI Insights
            ai_prompt = f"Evaluate the effectiveness of {', '.join(programs)} based on {', '.join(metrics)}. Provide a detailed analysis of each program's performance, strengths, weaknesses, and recommendations for improvement."
            ai_insights = call_ai_model(ai_prompt)
            st.markdown("### AI-Powered Evaluation Insights")
            st.write(ai_insights)
            
            # Impact Visualization
            impact_data = {prog: random.uniform(0, 1) for prog in programs}
            fig = px.pie(values=list(impact_data.values()), names=list(impact_data.keys()), title="Program Impact Distribution")
            st.plotly_chart(fig)
            
            # Comparative Analysis
            st.markdown("### Comparative Analysis")
            comp_data = pd.DataFrame({metric: [random.uniform(0, 100) for _ in programs] for metric in metrics}, index=programs)
            fig = px.imshow(comp_data, text_auto=True, aspect="auto", title="Program Performance Heatmap")
            st.plotly_chart(fig)
    
    st.markdown('</div>', unsafe_allow_html=True)

elif page == "Barrier Identification":
    animated_progress_bar()
    st.markdown('<div class="section">', unsafe_allow_html=True)
    st.subheader("Digital Skills Acquisition Barriers")
    
    barrier_categories = ["Access to Technology", "Digital Literacy", "Financial Constraints", "Cultural Factors", "Education System", "Gender Disparity", "Infrastructure", "Language Barriers"]
    selected_barriers = st.multiselect("Select barrier categories to analyze:", barrier_categories)
    
    if st.button("Identify Barriers"):
        with st.spinner("Analyzing barriers..."):
            # Simulated data generation
            barrier_data = generate_simulated_data(selected_barriers)

            # Plotting
            fig = go.Figure(data=[go.Scatterpolar(
                r=[max(barrier_data[cat]) for cat in selected_barriers],
                theta=selected_barriers,
                fill='toself'
            )])
            fig.update_layout(
                polar=dict(
                    radialaxis=dict(visible=True),
                ),
                showlegend=False
            )
            fig.update_layout(title="Barrier Intensity Analysis")
            st.plotly_chart(fig)
            
            # AI Insights
            ai_prompt = f"with a precise answer, not so long and well structured for easy understanding. Analyze the barriers to digital skills acquisition in Kenya, focusing on {', '.join(selected_barriers)}. Provide a comprehensive breakdown of each barrier, its impact, and potential solutions."
            ai_insights = call_ai_model(ai_prompt)
            st.markdown("### AI-Generated Barrier Analysis")
            st.write(ai_insights)
            
            # Recommendations
            st.markdown("### Recommended Interventions")
            interventions = {
                "Access to Technology": "Implement mobile learning programs and community technology centers",
                "Digital Literacy": "Integrate digital skills into primary and secondary education curricula",
                "Financial Constraints": "Provide scholarships and low-interest loans for digital skills training",
                "Cultural Factors": "Develop culturally sensitive training materials and awareness campaigns",
                "Education System": "Partner with local schools for early digital education integration",
                "Gender Disparity": "Create women-focused digital skills programs and mentorship opportunities",
                "Infrastructure": "Invest in rural broadband expansion and solar-powered internet kiosks",
                "Language Barriers": "Develop multilingual digital learning resources and interfaces"
            }
            for barrier in selected_barriers:
                st.write(f"- **{barrier}:** {interventions.get(barrier, 'Customized intervention based on specific barrier characteristics')}")
            
            # Barrier Impact Over Time
            st.markdown("### Projected Barrier Impact Over Time")
            years = [datetime.now().year + i for i in range(5)]
            impact_data = {barrier: [random.uniform(0, 100) for _ in range(5)] for barrier in selected_barriers}
            fig = go.Figure()
            for barrier, impact in impact_data.items():
                fig.add_trace(go.Scatter(x=years, y=impact, mode='lines+markers', name=barrier))
            fig.update_layout(title="Projected Barrier Impact (Next 5 Years)", xaxis_title="Year", yaxis_title="Impact Score")
            st.plotly_chart(fig)
    
    st.markdown('</div>', unsafe_allow_html=True)

elif page == "Recommendations":
    animated_progress_bar()
    st.markdown('<div class="section">', unsafe_allow_html=True)
    st.subheader("Personalized Recommendations")
    
    # User profile input
    st.write("Enter your profile information to receive personalized recommendations:")
    age = st.slider("Age", 15, 35, 25)
    education = st.selectbox("Highest Education Level", ["High School", "Diploma", "Bachelor's", "Master's", "PhD"])
    current_skills = st.multiselect("Current Digital Skills", 
                                    ["Basic Computer Use", "Office Software", "Web Browsing", "Social Media", 
                                     "Programming", "Data Analysis", "Graphic Design", "Digital Marketing"])
    career_goal = st.text_input("Career Goal")
    location = st.selectbox("Current Location", ["Nairobi", "Mombasa", "Kisumu", "Nakuru", "Eldoret", "Kisii", "Nyamira"])
    
    if st.button("Get Recommendations"):
        with st.spinner("Generating personalized recommendations..."):
            # AI-generated recommendations
            ai_prompt = f"""Generate personalized digital skills recommendations for a {age}-year-old with {education} education, 
            skills in {', '.join(current_skills)}, located in {location}, aiming for a career in {career_goal}. 
            Provide a detailed learning path, skill priorities, job market insights, and potential challenges to overcome. Well structured for easy understanding"""
            recommendations = call_ai_model(ai_prompt)
            
            st.markdown("### Your Personalized Digital Skills Roadmap")
            st.write(recommendations)
            
            # Simulated skill gap analysis
            st.markdown("### Skill Gap Analysis")
            all_skills = ["Data Analysis", "Web Development", "Digital Marketing", "Cybersecurity", "Cloud Computing", 
                          "AI/Machine Learning", "Mobile App Development", "UI/UX Design"]
            skill_gaps = [skill for skill in all_skills if skill not in current_skills]
            gap_scores = [random.randint(1, 100) for _ in skill_gaps]
            fig = go.Figure(go.Bar(
                x=skill_gaps,
                y=gap_scores,
                marker_color=['#1e3799', '#4a69bd', '#6a89cc', '#54a0ff', '#48dbfb']
            ))
            fig.update_layout(title="Skill Gap Analysis", xaxis_title="Skills", yaxis_title="Proficiency Gap")
            st.plotly_chart(fig)
            
            # Job market alignment
            st.markdown("### Job Market Alignment")
            job_matches = ["Data Analyst", "Digital Marketing Specialist", "Front-end Developer", "UX Researcher", "Cybersecurity Analyst"]
            match_scores = [random.uniform(0.5, 1) for _ in job_matches]
            fig = px.bar(x=job_matches, y=match_scores, labels={'x': 'Job Titles', 'y': 'Match Score'}, title="Top Job Matches")
            st.plotly_chart(fig)
            
            # Learning resources
            st.markdown("### Recommended Learning Resources")
            resources = [
                {"name": "Coursera", "type": "Online Courses", "link": "https://www.coursera.org/"},
                {"name": "edX", "type": "Online Courses", "link": "https://www.edx.org/"},
                {"name": "Udacity", "type": "Nanodegree Programs", "link": "https://www.udacity.com/"},
                {"name": "FreeCodeCamp", "type": "Coding Tutorials", "link": "https://www.freecodecamp.org/"},
                {"name": "Codecademy", "type": "Interactive Coding Lessons", "link": "https://www.codecademy.com/"}
            ]
            for resource in resources:
                st.write(f"- [{resource['name']}]({resource['link']}) - {resource['type']}")
    
    st.markdown('</div>', unsafe_allow_html=True)

elif page == "Reports":
    animated_progress_bar()
    st.markdown('<div class="section">', unsafe_allow_html=True)
    st.subheader("Generate Custom Reports")
    
    report_type = st.selectbox("Select Report Type", ["Skills Demand", "Program Effectiveness", "Barrier Analysis"])
    date_range = st.date_input("Select Date Range", [datetime.now() - timedelta(days=30), datetime.now()])
    
    if st.button("Generate Report"):
        with st.spinner("Generating report..."):
            # Simulated report generation
            if report_type == "Skills Demand":
                skills = ["Data Analysis", "Web Development", "Digital Marketing", "Cybersecurity", "Cloud Computing"]
                demand_data = pd.DataFrame({
                    "Skill": skills,
                    "Demand Score": [random.randint(50, 100) for _ in skills],
                    "Growth Rate": [random.uniform(0.5, 15) for _ in skills]
                })
                
                st.markdown("### Skills Demand Report")
                st.dataframe(demand_data)
                
                fig = px.scatter(demand_data, x="Demand Score", y="Growth Rate", size="Demand Score", 
                                 color="Skill", hover_name="Skill", size_max=60)
                fig.update_layout(title="Skills Demand vs Growth Rate")
                st.plotly_chart(fig)
                
            elif report_type == "Program Effectiveness":
                programs = ["Ajira Digital", "DigiTruck", "eMobilis", "KamiLimu", "Andela"]
                metrics = ["Employment Rate", "Income Increase", "Skill Proficiency", "Program Completion Rate"]
                effectiveness_data = pd.DataFrame({
                    metric: [random.uniform(60, 95) for _ in programs] for metric in metrics
                }, index=programs)
                
                st.markdown("### Program Effectiveness Report")
                st.dataframe(effectiveness_data)
                
                fig = px.imshow(effectiveness_data, text_auto=True, aspect="auto", 
                                title="Program Effectiveness Heatmap")
                st.plotly_chart(fig)
                
            elif report_type == "Barrier Analysis":
                barriers = ["Access to Technology", "Digital Literacy", "Financial Constraints", "Cultural Factors"]
                impact_data = pd.DataFrame({
                    "Barrier": barriers,
                    "Impact Score": [random.uniform(3, 9) for _ in barriers],
                    "Affected Population (%)": [random.uniform(10, 50) for _ in barriers]
                })
                
                st.markdown("### Barrier Analysis Report")
                st.dataframe(impact_data)
                
                fig = px.bar(impact_data, x="Barrier", y="Impact Score", color="Affected Population (%)",
                             title="Barrier Impact Analysis")
                st.plotly_chart(fig)
            
            # Generate Excel report
            excel_data = {
                "Report": pd.DataFrame({"Report Type": [report_type], "Date Range": [f"{date_range[0]} to {date_range[1]}"]})
            }
            if report_type == "Skills Demand":
                excel_data["Skills Demand"] = demand_data
            elif report_type == "Program Effectiveness":
                excel_data["Program Effectiveness"] = effectiveness_data
            elif report_type == "Barrier Analysis":
                excel_data["Barrier Analysis"] = impact_data
            
            excel_report = create_excel_report(excel_data)
            st.download_button(
                label="Download Excel Report",
                data=excel_report,
                file_name=f"DSEAP_{report_type}_Report_{date_range[0]}_{date_range[1]}.xlsx",
                mime="application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
            )
    
    st.markdown('</div>', unsafe_allow_html=True)

# Footer
st.markdown('<div class="footer">© 2024 Digital Skills and Employment Analytics Platform (DSEAP) </div>', unsafe_allow_html=True)

# Add a warning about simulated data
st.markdown('<div class="warning-box">', unsafe_allow_html=True)
st.warning("**Note:** This is a prototype version Developed by Teresa Abuya, COD SIST")
st.markdown('</div>', unsafe_allow_html=True)