phjacoby/ETL_Datamanagement
1
1
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

finishes setup

Browse Source
This commit is contained in:
Philipp Jacoby
2026-02-10 17:43:26 +01:00
parent 4f1a5c311f
commit 3003310be0
39 changed files with 2251611 additions and 1188 deletions
Download Patch File Download Diff File Expand all files Collapse all files

559
dashboard.py Normal file
View File

@@ -0,0 +1,559 @@
import streamlit as st
import pandas as pd
import plotly.express as px
import plotly.graph_objects as go
import networkx as nx
from pathlib import Path
st.set_page_config(page_title="Hetionet Dashboard", layout="wide")
# Custom CSS
st.markdown("""
<style>
.stTabs [data-baseweb="tab-list"] {
gap: 24px;
}
.stTabs [data-baseweb="tab"] {
height: 50px;
padding-left: 20px;
padding-right: 20px;
}
.stat-box {
background: grey;
padding: 20px;
border-radius: 10px;
color: white;
text-align: center;
margin: 10px 0;
}
</style>
""", unsafe_allow_html=True)
st.title("Hetionet Drug Analysis Dashboard")
# Sidebar
st.sidebar.header("Global Search")
search_type = st.sidebar.selectbox("Search for:", ["Gene", "Disease", "Drug"])
search_term = st.sidebar.text_input(f"Enter {search_type} name:")
st.sidebar.markdown("---")
st.sidebar.header("Navigation")
page = st.sidebar.radio("Select Analysis:", [
"Overview",
"Hotspot Genes",
"Drug Repurposing",
"Polypharmacy Risk",
"Symptom Triangle",
"Super Drugs",
"Drug Conflicts",
"Network Graph",
"Compare Drugs"
])
# Load data
try:
data_dir = Path("neo4j_csv")
with st.spinner("Loading data..."):
genes = pd.read_csv(data_dir / "nodes_Gene.csv")
diseases = pd.read_csv(data_dir / "nodes_Disease.csv")
repurposing = pd.read_csv(data_dir / "analysis_drug_repurposing.csv")
polypharmacy = pd.read_csv(data_dir / "analysis_polypharmacy_risk.csv")
symptom_triangle = pd.read_csv(data_dir / "analysis_symptom_triangle.csv")
super_drugs = pd.read_csv(data_dir / "analysis_super_drugs.csv")
# Try loading new files
try:
drug_conflicts = pd.read_csv(data_dir / "analysis_drug_conflicts.csv")
network_nodes = pd.read_csv(data_dir / "network_nodes.csv")
network_edges = pd.read_csv(data_dir / "network_edges.csv")
except:
drug_conflicts = None
network_nodes = None
network_edges = None
# Global Search Results
if search_term:
st.sidebar.markdown("---")
st.sidebar.subheader("Search Results")
if search_type == "Gene":
results = genes[genes['name'].str.contains(search_term, case=False, na=False)]
if len(results) > 0:
st.sidebar.success(f"Found {len(results)} genes")
for _, row in results.head(5).iterrows():
st.sidebar.write(f"**{row['name']}**: {row['num_diseases']} diseases")
else:
st.sidebar.warning("No genes found")
elif search_type == "Disease":
results = diseases[diseases['name'].str.contains(search_term, case=False, na=False)]
if len(results) > 0:
st.sidebar.success(f"Found {len(results)} diseases")
for _, row in results.head(5).iterrows():
st.sidebar.write(f"**{row['name']}**: {row['num_symptoms']} symptoms")
else:
st.sidebar.warning("No diseases found")
elif search_type == "Drug":
results = super_drugs[super_drugs['name'].str.contains(search_term, case=False, na=False)]
if len(results) > 0:
st.sidebar.success(f"Found {len(results)} drugs")
for _, row in results.head(5).iterrows():
st.sidebar.write(f"**{row['name']}**: Score {row['super_score']:.2f}")
else:
st.sidebar.warning("No drugs found")
# OVERVIEW PAGE
if page == "Overview":
st.header("Dataset Overview")
col1, col2, col3, col4 = st.columns(4)
col1.metric("Total Genes", f"{len(genes):,}")
col2.metric("Total Diseases", f"{len(diseases):,}")
col3.metric("Repurposing Opportunities", f"{len(repurposing):,}")
col4.metric("Analyzed Drugs", f"{len(super_drugs):,}")
# STATISTICS BOXES
st.markdown("---")
st.subheader("Key Statistics")
col1, col2, col3 = st.columns(3)
with col1:
avg_diseases_per_gene = genes[genes['num_diseases'] > 0]['num_diseases'].mean()
st.markdown(f"""
<div class="stat-box">
<h3>avg. Diseases per Gene</h3>
<h1>{avg_diseases_per_gene:.1f}</h1>
<p>For genes with disease associations</p>
</div>
""", unsafe_allow_html=True)
with col2:
avg_symptoms_per_disease = diseases[diseases['num_symptoms'] > 0]['num_symptoms'].mean()
st.markdown(f"""
<div class="stat-box">
<h3>avg. symptoms per disease</h3>
<h1>{avg_symptoms_per_disease:.1f}</h1>
<p>For diseases with documented symptoms</p>
</div>
""", unsafe_allow_html=True)
with col3:
avg_side_effects = polypharmacy['num_side_effects'].mean()
st.markdown(f"""
<div class="stat-box">
<h3>avg. side effect per drug</h3>
<h1>{avg_side_effects:.1f}</h1>
<p>Across all analyzed compounds</p>
</div>
""", unsafe_allow_html=True)
st.markdown("---")
col1, col2 = st.columns(2)
with col1:
st.subheader("Top Genes by Disease Count")
top_genes = genes.nlargest(10, 'num_diseases')[['name', 'num_diseases']]
fig = px.bar(top_genes, x='name', y='num_diseases', color='num_diseases')
fig.update_layout(
showlegend=False,
xaxis_title="Gene",
yaxis_title="Number of Diseases"
)
# Enable chart export
config = {'displayModeBar': True, 'displaylogo': False}
st.plotly_chart(fig, use_container_width=True, config=config)
csv = top_genes.to_csv(index=False).encode('utf-8')
st.download_button("Download Data", csv, "top_genes.csv", "text/csv")
with col2:
st.subheader("Top Diseases by Symptom Count")
top_diseases = diseases.nlargest(10, 'num_symptoms')[['name', 'num_symptoms']]
fig = px.bar(top_diseases, x='name', y='num_symptoms', color='num_symptoms', color_continuous_scale='Reds')
fig.update_layout(
showlegend=False,
xaxis_title="Disease",
yaxis_title="Number of Symptoms"
)
config = {'displayModeBar': True, 'displaylogo': False}
st.plotly_chart(fig, use_container_width=True, config=config)
csv = top_diseases.to_csv(index=False).encode('utf-8')
st.download_button("Download Data", csv, "top_diseases.csv", "text/csv")
# HOTSPOT GENES PAGE
elif page == "Hotspot Genes":
st.header("Hotspot Genes - Most Disease Associations")
col1, col2 = st.columns([3, 1])
with col2:
n_genes = st.slider("Number of genes:", 10, 50, 20)
min_diseases = st.slider("Min diseases:", 0, 50, 0)
with col1:
filtered_genes = genes[genes['num_diseases'] >= min_diseases].nlargest(n_genes, 'num_diseases')
fig = px.bar(
filtered_genes,
x='name',
y='num_diseases',
title=f'Top {n_genes} Genes (min {min_diseases} diseases)',
color='num_diseases',
color_continuous_scale='Viridis'
)
fig.update_layout(height=600, showlegend=False)
config = {'displayModeBar': True, 'displaylogo': False}
st.plotly_chart(fig, use_container_width=True, config=config)
st.dataframe(filtered_genes[['name', 'num_diseases']].reset_index(drop=True), use_container_width=True)
csv = filtered_genes.to_csv(index=False).encode('utf-8')
st.download_button("Download Filtered Data", csv, "hotspot_genes.csv", "text/csv")
# DRUG REPURPOSING PAGE
elif page == "Drug Repurposing":
st.header("Drug Repurposing Opportunities")
col1, col2 = st.columns([2, 1])
with col1:
top_n = st.slider("Show top N opportunities:", 10, 50, 20)
top_repurpose = repurposing.nlargest(top_n, 'shared_genes')
fig = px.scatter(
top_repurpose,
x='disease',
y='candidate_drug',
size='shared_genes',
color='shared_genes',
title='Drug Repurposing Candidates',
color_continuous_scale='Reds',
height=600
)
fig.update_xaxes(tickangle=-45)
config = {'displayModeBar': True, 'displaylogo': False}
st.plotly_chart(fig, use_container_width=True, config=config)
with col2:
st.subheader("Filter by Disease")
diseases_list = sorted(repurposing['disease'].unique())
selected = st.selectbox("Select disease:", diseases_list)
filtered = repurposing[repurposing['disease'] == selected].nlargest(10, 'shared_genes')
st.dataframe(filtered[['candidate_drug', 'shared_genes']].reset_index(drop=True), height=400)
csv = filtered.to_csv(index=False).encode('utf-8')
st.download_button("Download", csv, f"repurposing_{selected}.csv", "text/csv")
# POLYPHARMACY RISK PAGE
elif page == "Polypharmacy Risk":
st.header("Polypharmacy Risk Analysis")
col1, col2 = st.columns([3, 1])
with col2:
top_n = st.slider("Number of drugs:", 10, 30, 20)
min_risk = st.slider("Min risk score:", 0, 100, 0)
with col1:
filtered_risk = polypharmacy[polypharmacy['risk_score'] >= min_risk].nlargest(top_n, 'num_side_effects')
fig = px.scatter(
filtered_risk,
x='num_diseases_treated',
y='num_side_effects',
size='risk_score',
color='risk_score',
hover_data=['name'],
title='Drugs: Side Effects vs Diseases Treated',
color_continuous_scale='Reds'
)
fig.update_layout(height=600)
config = {'displayModeBar': True, 'displaylogo': False}
st.plotly_chart(fig, use_container_width=True, config=config)
st.dataframe(
filtered_risk[['name', 'num_diseases_treated', 'num_side_effects', 'risk_score']].reset_index(drop=True),
use_container_width=True
)
csv = filtered_risk.to_csv(index=False).encode('utf-8')
st.download_button("Download Risk Data", csv, "polypharmacy_risk.csv", "text/csv")
# SYMPTOM TRIANGLE PAGE
elif page == "Symptom Triangle":
st.header("Symptom-Disease-Drug Connections")
top_n = st.slider("Number of symptoms:", 10, 30, 20)
top_symptoms = symptom_triangle.nlargest(top_n, 'impact_score')
fig = px.scatter(
top_symptoms,
x='num_diseases',
y='num_treating_drugs',
size='impact_score',
color='drugs_with_side_effects',
hover_data=['symptom'],
title='Symptom Impact Analysis',
color_continuous_scale='RdYlGn_r'
)
fig.update_layout(height=600)
config = {'displayModeBar': True, 'displaylogo': False}
st.plotly_chart(fig, use_container_width=True, config=config)
st.dataframe(
top_symptoms[['symptom', 'num_diseases', 'num_treating_drugs', 'drugs_with_side_effects', 'impact_score']].reset_index(drop=True),
use_container_width=True
)
csv = top_symptoms.to_csv(index=False).encode('utf-8')
st.download_button("Download Symptom Data", csv, "symptom_triangle.csv", "text/csv")
# SUPER DRUGS PAGE
elif page == "Super Drugs":
st.header("Super-Drug Score (Best Benefit/Risk Ratio)")
col1, col2 = st.columns([3, 1])
with col2:
top_n = st.slider("Number of drugs:", 10, 30, 20)
min_score = st.slider("Min super score:", 0.0, 5.0, 0.0, 0.1)
with col1:
filtered_super = super_drugs[super_drugs['super_score'] >= min_score].nlargest(top_n, 'super_score')
fig = px.scatter(
filtered_super,
x='num_diseases_treated',
y='num_side_effects',
size='super_score',
color='super_score',
hover_data=['name'],
title='Super Drugs Analysis',
color_continuous_scale='Viridis_r'
)
fig.update_layout(height=600)
config = {'displayModeBar': True, 'displaylogo': False}
st.plotly_chart(fig, use_container_width=True, config=config)
st.dataframe(
filtered_super[['name', 'num_diseases_treated', 'num_side_effects', 'super_score']].reset_index(drop=True),
use_container_width=True
)
perfect = super_drugs[(super_drugs['num_side_effects'] == 0) & (super_drugs['num_diseases_treated'] > 0)]
st.info(f"Found {len(perfect)} drugs with ZERO documented side effects!")
csv = filtered_super.to_csv(index=False).encode('utf-8')
st.download_button("Download Super Drugs", csv, "super_drugs.csv", "text/csv")
# DRUG CONFLICTS PAGE
elif page == "Drug Conflicts":
st.header("Drug Conflicts - Overlapping Side Effects")
if drug_conflicts is not None and len(drug_conflicts) > 0:
col1, col2 = st.columns([3, 1])
with col2:
top_n = st.slider("Number of conflicts:", 10, 50, 20)
min_overlap = st.slider("Min shared side effects:", 0, 100, 10)
with col1:
filtered_conflicts = drug_conflicts[
drug_conflicts['shared_side_effects'] >= min_overlap
].nlargest(top_n, 'shared_side_effects')
fig = px.scatter(
filtered_conflicts,
x='drug1_total_se',
y='drug2_total_se',
size='shared_side_effects',
color='overlap_percentage',
hover_data=['drug1', 'drug2', 'shared_side_effects'],
title='Drug Pairs with Overlapping Side Effects',
labels={'drug1_total_se': 'Drug 1 Total SE', 'drug2_total_se': 'Drug 2 Total SE'},
color_continuous_scale='Reds'
)
fig.update_layout(height=600)
config = {'displayModeBar': True, 'displaylogo': False}
st.plotly_chart(fig, use_container_width=True, config=config)
st.warning("These drug combinations may have compounded side effects!")
st.dataframe(
filtered_conflicts[['drug1', 'drug2', 'shared_side_effects', 'overlap_percentage']].reset_index(drop=True),
use_container_width=True
)
csv = filtered_conflicts.to_csv(index=False).encode('utf-8')
st.download_button("Download Conflicts", csv, "drug_conflicts.csv", "text/csv")
else:
st.warning("Drug conflicts data not available. Run the ETL script to generate this analysis.")
# NETWORK GRAPH PAGE
elif page == "Network Graph":
st.header("Disease-Gene-Drug Network")
if network_nodes is not None and network_edges is not None:
st.info("Interactive network visualization showing connections between diseases, genes, and drugs")
# Create networkx graph
G = nx.Graph()
# Add nodes
for _, row in network_nodes.iterrows():
G.add_node(row['id'], label=row['label'], type=row['type'])
# Add edges
for _, row in network_edges.iterrows():
G.add_edge(row['source'], row['target'])
# Create layout
pos = nx.spring_layout(G, k=0.5, iterations=50)
# Create edge trace
edge_x = []
edge_y = []
for edge in G.edges():
x0, y0 = pos[edge[0]]
x1, y1 = pos[edge[1]]
edge_x.extend([x0, x1, None])
edge_y.extend([y0, y1, None])
edge_trace = go.Scatter(
x=edge_x, y=edge_y,
line=dict(width=0.5, color='#888'),
hoverinfo='none',
mode='lines'
)
# Create node traces (separate by type for legend)
node_traces = []
color_map = {
'Disease': '#ff4444',
'Gene': '#4444ff',
'Compound': '#44ff44'
}
for node_type, color in color_map.items():
node_x = []
node_y = []
node_text = []
for node in G.nodes():
if G.nodes[node]['type'] == node_type:
x, y = pos[node]
node_x.append(x)
node_y.append(y)
node_text.append(f"{node_type}: {G.nodes[node]['label']}")
if node_x:
node_trace = go.Scatter(
x=node_x, y=node_y,
mode='markers',
name=node_type,
hoverinfo='text',
text=node_text,
marker=dict(
color=color,
size=12,
line=dict(width=2, color='white')
)
)
node_traces.append(node_trace)
# Create figure
fig = go.Figure(data=[edge_trace] + node_traces,
layout=go.Layout(
title='Disease-Gene-Drug Network',
showlegend=True,
hovermode='closest',
margin=dict(b=0,l=0,r=0,t=40),
xaxis=dict(showgrid=False, zeroline=False, showticklabels=False),
yaxis=dict(showgrid=False, zeroline=False, showticklabels=False),
height=700
))
config = {'displayModeBar': True, 'displaylogo': False}
st.plotly_chart(fig, use_container_width=True, config=config)
col1, col2, col3 = st.columns(3)
col1.metric("🔴 Diseases", len([n for n in G.nodes() if G.nodes[n]['type'] == 'Disease']))
col2.metric("🔵 Genes", len([n for n in G.nodes() if G.nodes[n]['type'] == 'Gene']))
col3.metric("🟢 Drugs", len([n for n in G.nodes() if G.nodes[n]['type'] == 'Compound']))
else:
st.warning("Network data not available. Run the ETL script to generate this visualization.")
# COMPARE DRUGS PAGE
elif page == "Compare Drugs":
st.header("⚖️ Compare Drugs Side-by-Side")
drug_names = sorted(super_drugs['name'].unique())
col1, col2 = st.columns(2)
with col1:
drug1 = st.selectbox("Select Drug 1:", drug_names, key='drug1')
with col2:
drug2 = st.selectbox("Select Drug 2:", drug_names, key='drug2')
if drug1 and drug2:
drug1_data = super_drugs[super_drugs['name'] == drug1].iloc[0]
drug2_data = super_drugs[super_drugs['name'] == drug2].iloc[0]
st.markdown("---")
col1, col2 = st.columns(2)
with col1:
st.subheader(f"{drug1}")
st.metric("Diseases Treated", int(drug1_data['num_diseases_treated']))
st.metric("Side Effects", int(drug1_data['num_side_effects']))
st.metric("Super Score", f"{drug1_data['super_score']:.2f}")
with col2:
st.subheader(f"{drug2}")
st.metric("Diseases Treated", int(drug2_data['num_diseases_treated']))
st.metric("Side Effects", int(drug2_data['num_side_effects']))
st.metric("Super Score", f"{drug2_data['super_score']:.2f}")
# Comparison chart
comparison_df = pd.DataFrame({
'Metric': ['Diseases Treated', 'Side Effects', 'Super Score'],
drug1: [drug1_data['num_diseases_treated'], drug1_data['num_side_effects'], drug1_data['super_score']],
drug2: [drug2_data['num_diseases_treated'], drug2_data['num_side_effects'], drug2_data['super_score']]
})
fig = px.bar(
comparison_df,
x='Metric',
y=[drug1, drug2],
barmode='group',
title='Side-by-Side Comparison'
)
config = {'displayModeBar': True, 'displaylogo': False}
st.plotly_chart(fig, use_container_width=True, config=config)
# Winner determination
st.markdown("---")
st.subheader("Recommendation")
if drug1_data['super_score'] > drug2_data['super_score']:
st.success(f"**{drug1}** has a better benefit/risk ratio (Super Score: {drug1_data['super_score']:.2f})")
elif drug2_data['super_score'] > drug1_data['super_score']:
st.success(f"**{drug2}** has a better benefit/risk ratio (Super Score: {drug2_data['super_score']:.2f})")
else:
st.info("Both drugs have the same Super Score")
except FileNotFoundError as e:
st.error(f"Could not find data files")
st.write("Please ensure you're running the dashboard from the correct directory with access to `neo4j_csv/` folder")
except Exception as e:
st.error(f"Error: {str(e)}")
st.exception(e)