| import pandas as pd |
| import plotly.express as px |
| import plotly.graph_objects as go |
| from plotly.subplots import make_subplots |
| from collections import Counter |
|
|
| def create_world_map(docs_df): |
| """Create interactive world map showing study distribution""" |
| if docs_df.empty: |
| return None |
| |
| |
| country_counts = Counter() |
| for countries_str in docs_df['study_countries'].dropna(): |
| if isinstance(countries_str, str) and countries_str.lower() != 'nan': |
| |
| countries = [c.strip() for c in countries_str.split(',')] |
| for country in countries: |
| country_counts[country] += 1 |
| |
| if not country_counts: |
| return None |
| |
| |
| countries = list(country_counts.keys()) |
| counts = list(country_counts.values()) |
| |
| fig = go.Figure(data=go.Choropleth( |
| locations=countries, |
| z=counts, |
| locationmode='country names', |
| colorscale='Viridis', |
| text=countries, |
| hovertemplate='<b>%{text}</b><br>Studies: %{z}<extra></extra>', |
| colorbar_title="Number of Studies" |
| )) |
| |
| fig.update_layout( |
| title={ |
| 'text': '🌍 Global Research Coverage', |
| 'x': 0.5, |
| 'font': {'size': 20} |
| }, |
| geo=dict( |
| showframe=False, |
| showcoastlines=True, |
| projection_type='equirectangular' |
| ), |
| height=500 |
| ) |
| |
| return fig |
|
|
| def create_sector_analysis(docs_df): |
| """Create sector distribution charts""" |
| if docs_df.empty: |
| return None, None |
| |
| |
| sector_counts = docs_df['world_bank_sector'].value_counts().head(10) |
| |
| fig1 = px.bar( |
| x=sector_counts.values, |
| y=sector_counts.index, |
| orientation='h', |
| title="📊 Research by World Bank Sector", |
| labels={'x': 'Number of Studies', 'y': 'Sector'}, |
| color=sector_counts.values, |
| color_continuous_scale='viridis' |
| ) |
| fig1.update_layout(height=400, showlegend=False) |
| |
| |
| design_counts = docs_df['research_design'].value_counts().head(8) |
| |
| fig2 = px.pie( |
| values=design_counts.values, |
| names=design_counts.index, |
| title="🔬 Research Design Distribution", |
| color_discrete_sequence=px.colors.qualitative.Set3 |
| ) |
| fig2.update_traces(textposition='inside', textinfo='percent+label') |
| fig2.update_layout(height=400) |
| |
| return fig1, fig2 |
|
|
| def create_methodology_dashboard(docs_df): |
| """Create methodology analysis dashboard""" |
| if docs_df.empty: |
| return None |
| |
| |
| fig = make_subplots( |
| rows=2, cols=2, |
| subplot_titles=('Sample Size Distribution', 'Rigor Scores', |
| 'Data Collection Methods', 'Quality Indicators'), |
| specs=[[{"secondary_y": False}, {"secondary_y": False}], |
| [{"secondary_y": False}, {"secondary_y": False}]] |
| ) |
| |
| |
| sample_sizes = pd.to_numeric(docs_df['sample_size'], errors='coerce').dropna() |
| if not sample_sizes.empty: |
| fig.add_trace( |
| go.Histogram(x=sample_sizes, name="Sample Size", nbinsx=20), |
| row=1, col=1 |
| ) |
| |
| |
| rigor_scores = pd.to_numeric(docs_df['rigor_score'], errors='coerce').dropna() |
| if not rigor_scores.empty: |
| fig.add_trace( |
| go.Histogram(x=rigor_scores, name="Rigor Score", nbinsx=10), |
| row=1, col=2 |
| ) |
| |
| |
| data_methods = docs_df['data_collection_method'].value_counts().head(8) |
| if not data_methods.empty: |
| fig.add_trace( |
| go.Bar(x=data_methods.values, y=data_methods.index, |
| orientation='h', name="Data Methods"), |
| row=2, col=1 |
| ) |
| |
| |
| quality_data = [] |
| for col in ['has_randomization', 'has_validation', 'has_mixed_methods']: |
| if col in docs_df.columns: |
| true_count = (docs_df[col] == 'true').sum() |
| quality_data.append((col.replace('has_', '').title(), true_count)) |
| |
| if quality_data: |
| labels, values = zip(*quality_data) |
| fig.add_trace( |
| go.Bar(x=list(labels), y=list(values), name="Quality Features"), |
| row=2, col=2 |
| ) |
| |
| fig.update_layout( |
| height=800, |
| title_text="📈 Methodology Dashboard", |
| title_x=0.5, |
| showlegend=False |
| ) |
| |
| return fig |
|
|
| def filter_studies(docs_df, countries, sectors, min_year, max_year, has_rct, min_sample_size): |
| """Filter and display studies based on criteria""" |
| if docs_df.empty: |
| return "No data available" |
| |
| filtered = docs_df.copy() |
| |
| |
| if countries: |
| country_mask = filtered['study_countries'].str.contains('|'.join(countries), case=False, na=False) |
| filtered = filtered[country_mask] |
| |
| if sectors: |
| sector_mask = filtered['world_bank_sector'].isin(sectors) |
| filtered = filtered[sector_mask] |
| |
| if min_year: |
| year_col = pd.to_numeric(filtered['publication_year'], errors='coerce') |
| filtered = filtered[year_col >= min_year] |
| |
| if max_year: |
| year_col = pd.to_numeric(filtered['publication_year'], errors='coerce') |
| filtered = filtered[year_col <= max_year] |
| |
| if has_rct: |
| filtered = filtered[filtered['has_randomization'] == 'true'] |
| |
| if min_sample_size: |
| sample_col = pd.to_numeric(filtered['sample_size'], errors='coerce') |
| filtered = filtered[sample_col >= min_sample_size] |
| |
| |
| if filtered.empty: |
| return "No studies match the selected criteria." |
| |
| |
| result = f"## 🔍 Filtered Results: {len(filtered)} studies\n\n" |
| |
| |
| display_cols = ['title', 'authors', 'publication_year', 'study_countries', |
| 'world_bank_sector', 'research_design', 'sample_size'] |
| available_cols = [col for col in display_cols if col in filtered.columns] |
| |
| sample_df = filtered[available_cols].head(10) |
| result += sample_df.to_markdown(index=False) |
| |
| if len(filtered) > 10: |
| result += f"\n\n*... and {len(filtered) - 10} more studies*" |
| |
| return result |
|
|
| def get_unique_values(docs_df): |
| """Extract unique countries and sectors for dropdowns""" |
| countries_list = [] |
| sectors_list = [] |
| |
| if not docs_df.empty: |
| |
| for countries_str in docs_df['study_countries'].dropna(): |
| if isinstance(countries_str, str) and countries_str.lower() != 'nan': |
| countries = [c.strip() for c in countries_str.split(',')] |
| countries_list.extend(countries) |
| countries_list = sorted(list(set(countries_list))) |
| |
| |
| sectors_list = sorted(docs_df['world_bank_sector'].dropna().unique().tolist()) |
| |
| return countries_list, sectors_list |
