add heatmap layer

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
 wealth_map.tif filter=lfs diff=lfs merge=lfs -text
+*.geojson filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -1,4 +1,3 @@
 .venv/
 __pycache__
-.DS_Store
-keys.py
+.DS_Store

app.py

@@ -6,11 +6,11 @@ import matplotlib.pyplot as plt
 import rasterio
 from rasterio.plot import show
 import geopandas as gpd
-from ipyleaflet import Map, TileLayer, basemaps, ColorMap, RasterLayer, LegendControl, GeoJSON
+from ipyleaflet import Map, TileLayer, basemaps, ColorMap, RasterLayer, LegendControl, GeoJSON, MarkerCluster, Marker, DivIcon, Polygon
 from shinywidgets import output_widget, register_widget
 import plotnine as p9
-# from palettable.colorbrewer.diverging import Spectral_10 
-# from palettable.colorbrewer.sequential import Blues_9, OrRd_9, PuBuGn_9, Reds_9
+import matplotlib.cm as cm
+import matplotlib.colors as colors
 import os
 import base64
 import tempfile
@@ -18,6 +18,11 @@ import json
 from datetime import datetime
 from helpers.fetch_data import fetch_data
 from helpers.residuals import get_residual_plot
+from helpers.reduce_precision import reduce_coordinate_precision
+from shapely.geometry import shape
+from shapely.geometry import mapping
+from shapely.geometry import Polygon as ShapelyPolygon
+import io
 
 # ------------------------------
 # 1. Data & Config
@@ -25,21 +30,38 @@ from helpers.residuals import get_residual_plot
 
 # Define time periods corresponding to each band in the GeoTIFF
 time_periods = ["1990–1992", "1993–1995", "1996–1998", "1999–2001", "2002–2004",
-               "2005–2007", "2008–2010", "2011–2013", "2014–2016", "2017–2019"]
+                "2005–2007", "2008–2010", "2011–2013", "2014–2016", "2017–2019"]
 
 # Load GeoTIFF data (multi-band)
-# Note: In a real application, you'd need to adjust this path
 wealth_stack = rasterio.open("wealth_map.tif")
 
-with open('data/no_somaliland.geojson') as a: 
+#load country data
+with open('data/no_somaliland.geojson') as a:
     country_json = json.load(a)
-  
+
+#load IWI by country
 IWI_df = pd.read_csv('data/mean_IWI_by_country.csv')
 
+#load residual data
 residual_data = pd.read_csv('data/residual_by_country.csv')
 
+selected_map = reactive.Value(None)
+
+#load band 1 data by default
+with open('data/band_1.geojson') as w:
+    band_1_data = json.load(w)
+    band_1_data = reduce_coordinate_precision(band_1_data, precision=5) #reduce precision to aid in rendering 
+
+IWI_values = [feature['properties']['IWI']
+              for feature in band_1_data['features']]
+norm = colors.Normalize(vmin=min(IWI_values), vmax=max(IWI_values))
+colormap = cm.get_cmap("plasma")
 
 
+def get_color(iwi):
+    rgba = colormap(norm(iwi))  # Convert to RGBA
+    return colors.to_hex(rgba)  # Convert to HEX
+
 
 # Function to clean up out-of-range values and get values
 def get_clean_values(src, band_idx=1):
@@ -48,6 +70,7 @@ def get_clean_values(src, band_idx=1):
     band_data[(band_data <= 0) | (band_data > 1)] = np.nan
     return band_data
 
+
 # Get all values across all bands for quantiles
 all_vals = []
 for i in range(1, wealth_stack.count + 1):
@@ -59,7 +82,7 @@ q_breaks_legend = np.quantile(all_vals, np.linspace(0, 1, 6))
 q_breaks = np.quantile(all_vals, np.linspace(0, 1, 11))
 
 # Get raster bounds for proper positioning on the map
-bounds = [[wealth_stack.bounds.bottom, wealth_stack.bounds.left], 
+bounds = [[wealth_stack.bounds.bottom, wealth_stack.bounds.left],
           [wealth_stack.bounds.top, wealth_stack.bounds.right]]
 
 # Load improvement data (change in IWI by state/province)
@@ -225,106 +248,129 @@ app_ui = ui.page_fluid(
     ),
     ui.page_navbar(
         ui.nav_panel("Wealth Map",
-            ui.layout_sidebar(
-                ui.sidebar(
-                    ui.h4("Map Controls"),
-                    ui.input_slider(
-                        "time_index", 
-                        "Select Time Period (Years):", 
-                        min=1, 
-                        max=len(time_periods), 
-                        value=1, 
-                        step=1,
-                        animate=True
-                    ),
-                    ui.strong("Currently Selected: "),
-                    ui.output_text("current_year_range", inline=True),
-                    ui.input_select(
-                        "color_palette", 
-                        "Select Color Palette:",
-                        {
-                            "blue": "blue",
-                            "red": "red",
-                            "orange": "orange",
-                            "purple": "purple",
-                            "Spectral": "Spectral (Brewer)"
-                        },
-                        selected="red"
-                    ),
-                    ui.input_slider(
-                        "opacity", 
-                        "Map Opacity:", 
-                        min=0.2, 
-                        max=1, 
-                        value=0.8, 
-                        step=0.1
-                    ),
-                    ui.HTML(share_button_html)
-                ),
-                ui.layout_column_wrap(
-                    ui.value_box(
-                        "Highest IWI", 
-                        ui.output_text("highest_iwi"),
-                        showcase=ui.tags.i(class_="fa fa-arrow-up"),
-                        theme="success"
-                    ),
-                    ui.value_box(
-                        "Lowest IWI", 
-                        ui.output_text("lowest_iwi"),
-                        showcase=ui.tags.i(class_="fa fa-arrow-down"),
-                        theme="danger"
-                    ),
-                    ui.value_box(
-                        "Average IWI", 
-                        ui.output_text("avg_iwi"),
-                        showcase=ui.tags.i(class_="fa fa-balance-scale"),
-                        theme="primary"
-                    ),
-                    width=1/3
-                ),
-                ui.layout_column_wrap(
-                    ui.card(
-                        ui.card_header(ui.h3("Wealth Map of Africa", class_="title-text")),
-                        output_widget("map"),
-                        ui.p("Click anywhere on the map to view the time-series of IWI for that specific location (shown below).")
-                    ),
-                    ui.card(
-                      ui.card_header(ui.h3("Time Series at Clicked Location", class_="subtitle-text")),
-                      ui.output_plot("clicked_ts_plot"),
-                      ui.p("Click on the map to see the full IWI time-series (1990–2019) for that location.")
-                  )
-                ),
-                
-                ui.card(
-                        ui.card_header(ui.h3("Ground Truth vs. Prediction Residual Distribution (Selected Country)", class_="subtitle-text")),
-                        ui.output_plot("iwi_residuals"),
-                        ui.p("This chart shows the distribution of residuals between ground truth and predicted IWI values based on the selected country."),
-                        ui.strong("Note: wealth estimates for areas without human settlements have been excluded from the analysis."),
-                        ui.HTML("<a href='https://doi.org/10.24963/ijcai.2023/684' target='_blank'>[Paper PDF]</a>")
-                    ),
-            )
-        ),
+                     ui.layout_sidebar(
+                         ui.sidebar(
+                             ui.h4("Map Controls"),
+                             ui.input_switch(
+                                 "SelectedMap", "Enable Country View", False),
+                             ui.input_slider(
+                                 "time_index",
+                                 "Select Time Period (Years):",
+                                 min=1,
+                                 max=len(time_periods),
+                                 value=1,
+                                 step=1,
+                                 animate=True
+                             ),
+                             ui.strong("Currently Selected: "),
+                             ui.output_text("current_year_range", inline=True),
+                             ui.input_select(
+                                 "color_palette",
+                                 "Select Color Palette:",
+                                 {
+                                     "blue": "blue",
+                                     "red": "red",
+                                     "orange": "orange",
+                                     "purple": "purple",
+                                     "Spectral": "Spectral (Brewer)"
+                                 },
+                                 selected="red"
+                             ),
+                             ui.input_slider(
+                                 "opacity",
+                                 "Map Opacity:",
+                                 min=0.2,
+                                 max=1,
+                                 value=0.8,
+                                 step=0.1
+                             ),
+                             ui.accordion(ui.accordion_panel(
+                                 'How it works', ui.HTML("<p>These wealth-index predictions are AI-generated by a"
+                                                         "sequence-aware neural network trained on 30 years of <em>Demographic and Health Surveys (DHS)</em> ground-truth data.</p"
+                                                         "<ul><li>🔍 57,100+ geo-referenced survey points from DHS</li> <li>⚙️ Multi-spectral satellite bands & raster-to-vector feature extraction</li><li>🎯 Calibrated & validated with held-out DHS clusters (1990–2019)</li></ul>")
+                             ), id="map_instructions", open=False, multiple=False),
+                             ui.HTML(share_button_html)
+                         ),
+                         ui.layout_column_wrap(
+                             ui.value_box(
+                                 "Highest IWI",
+                                 ui.output_text("highest_iwi"),
+                                 showcase=ui.tags.i(class_="fa fa-arrow-up"),
+                                 theme="success"
+                             ),
+                             ui.value_box(
+                                 "Lowest IWI",
+                                 ui.output_text("lowest_iwi"),
+                                 showcase=ui.tags.i(class_="fa fa-arrow-down"),
+                                 theme="danger"
+                             ),
+                             ui.value_box(
+                                 "Average IWI",
+                                 ui.output_text("avg_iwi"),
+                                 showcase=ui.tags.i(
+                                     class_="fa fa-balance-scale"),
+                                 theme="primary"
+                             ),
+                             width=1/3
+                         ),
+                         ui.layout_column_wrap(
+                             ui.card(
+                                 ui.card_header(
+                                     ui.h3("Wealth Map of Africa", class_="title-text")),
+                                 output_widget("country_map"),
+                                 ui.p(
+                                     "Click anywhere on the map to view the time-series of IWI for that specific location (shown below).")
+                             ),
+                             ui.card(
+                                 ui.card_header(
+                                     ui.h3("Time Series at Clicked Location", class_="subtitle-text")),
+                                 ui.output_plot("clicked_ts_plot"),
+                                 ui.p(
+                                     "Click on the map to see the full IWI time-series (1990–2019) for that location."),
+                                 ui.download_button(
+                                     "download_country_data", "Download CSV", icon="download"),
+
+                             )
+                         ),
+
+                         ui.card(
+                             ui.card_header(ui.h3(
+                                 "Ground Truth vs. Prediction Residual Distribution (Selected Country)", class_="subtitle-text")),
+                             ui.output_plot("iwi_residuals"),
+                             ui.p(
+                                 "This chart shows the distribution of residuals between ground truth and predicted IWI values based on the selected country."),
+                             ui.strong(
+                                 "Note: wealth estimates for areas without human settlements have been excluded from the analysis."),
+                             ui.HTML(
+                                 "<a href='https://doi.org/10.24963/ijcai.2023/684' target='_blank'>[Paper PDF]</a>")
+                         )
+                     )
+                     ),
         ui.nav_panel("Improvement Data",
-            ui.layout_columns(
-            ui.card(
-                ui.card_header(ui.h3("Poverty Improvement by State", class_="title-text")),
-                ui.p("This table shows the estimated improvement in mean IWI between 1990–1992 and 2017–2019 for each province in Africa. "
-                   "The 'Improvement' column indicates the change in IWI over this period. You can sort or filter the table, "
-                   "and use the download button to export the data."),
-                ui.download_button("download_data", "Download CSV", icon="download"),
-                ui.card(ui.output_data_frame("improvement_table")),
-                
-            )
-            )
-        ),
+                     ui.layout_columns(
+                         ui.card(
+                             ui.card_header(
+                                 ui.h3("Poverty Improvement by State", class_="title-text")),
+                             ui.p("This table shows the estimated improvement in mean IWI between 1990–1992 and 2017–2019 for each province in Africa. "
+                                  "The 'Improvement' column indicates the change in IWI over this period. You can sort or filter the table, "
+                                  "and use the download button to export the data."),
+                             ui.download_button(
+                                 "download_data", "Download CSV", icon="download"),
+                             ui.card(ui.output_data_frame(
+                                 "improvement_table")),
+
+                         )
+                     )
+                     ),
         ui.nav_panel("Trends Over Time",
-            ui.card(
-                ui.card_header(ui.h3("Average Wealth Index Across Africa Over Time", class_="title-text")),
-                ui.p("This chart aggregates the mean IWI across all of Africa in each of the ten time periods. "
-                   "It provides a high-level view of how wealth (as measured by IWI) has changed over time."),
-                ui.output_plot("trend_plot")
-            )
-        ),
+                     ui.card(
+                         ui.card_header(
+                             ui.h3("Average Wealth Index Across Africa Over Time", class_="title-text")),
+                         ui.p("This chart aggregates the mean IWI across all of Africa in each of the ten time periods. "
+                              "It provides a high-level view of how wealth (as measured by IWI) has changed over time."),
+                         ui.output_plot("trend_plot")
+                     )
+                     ),
         title=ui.HTML(
             "<span style='font-weight: 600; font-size: 16px;'>"
             "<a href='http://aidevlab.org' target='_blank' "
@@ -339,9 +385,27 @@ app_ui = ui.page_fluid(
 # ------------------------------
 # 3. Server logic
 # ------------------------------
+
+
 def server(input, output, session):
     # Initialize the map widget
     m = Map(center=(0, 20), zoom=3)
+    for feature in band_1_data["features"]:
+        iwi = feature["properties"]["IWI"]
+        feature["properties"]["style"] = {
+            "color": get_color(iwi),
+            "fillColor": get_color(iwi),  # Fill color based on IWI
+            "fillOpacity": 0.7,
+            "weight": 1
+        }
+
+    band_1_json = GeoJSON(data=band_1_data,
+                          style={'radius': 2, 'opacity': 0.8, 'weight': 1.9},
+                          point_style={'radius': 3},
+                          name='Release'
+                          )
+    m.add_layer(band_1_json)
+
     geo_json = GeoJSON(
         data=country_json,
         style={
@@ -351,189 +415,136 @@ def server(input, output, session):
             'color': 'white', 'dashArray': '0', 'fillOpacity': 0.5
         }
     )
-    m.add_layer(geo_json)
-
-    
-
     # Register the map widget with Shiny
-    map_widget = register_widget("map", m)
-    
+    map_widget = register_widget("country_map", m)
+
     # Store clicked point values
     clicked_point_vals = reactive.Value(None)
     selected_country = reactive.Value(None)
 
     admin_layer = reactive.Value(None)
     selected_admin = reactive.Value(None)
-    
+
     # Get the currently selected raster layer
     @reactive.Calc
     def selected_raster():
         band_idx = input.time_index()
         return get_clean_values(wealth_stack, band_idx)
-    
+
     # Display selected time period
     @output
     @render.text
     def current_year_range():
-        return time_periods[input.time_index() - 1]  # Adjust for 0-based indexing
-    
-    # Function to get color palette based on user selection
-    # @reactive.Calc
-    # def get_palette():
-    #     palette_name = input.color_palette()
-    #     if palette_name == "blue":
-    #         return Blues_9.hex_colors 
-    #     elif palette_name == "orange":
-    #         return OrRd_9.hex_colors
-    #     elif palette_name == "red":
-    #         return Reds_9.hex_colors
-    #     elif palette_name == "purple":
-    #         return PuBuGn_9.hex_colors
-    #     else:  # Spectral
-    #         return Spectral_10.hex_colors
-    
-    # Create a RasterLayer for the map
-    # @reactive.effect
+        # Adjust for 0-based indexing
+        return time_periods[input.time_index() - 1]
+
+    # Create a Country layer for the map
+    @reactive.effect
     # @reactive.event(input.time_index, input.color_palette, input.opacity)
-    # def _():
-    #     # Remove existing raster layers
-    #     for layer in m.layers:
-    #         if isinstance(layer, RasterLayer):
-    #             m.remove_layer(layer)
-        
-    #     # Get current raster data
-    #     raster_data = selected_raster()
-        
-    #     # Create a temporary GeoTIFF file
-    #     with tempfile.NamedTemporaryFile(suffix='.tif', delete=False) as tmp:
-    #         temp_path = tmp.name
-            
-    #         # Create a new GeoTIFF with the selected band
-    #         with rasterio.open(
-    #             temp_path,
-    #             'w',
-    #             driver='GTiff',
-    #             height=raster_data.shape[0],
-    #             width=raster_data.shape[1],
-    #             count=1,
-    #             dtype=raster_data.dtype,
-    #             crs=wealth_stack.crs,
-    #             transform=wealth_stack.transform,
-    #         ) as dst:
-    #             dst.write(raster_data, 1)
-        
-    #     # Create a ColorMap for the raster
-    #     colormap = ColorMap(
-    #         vmin=q_breaks[0],
-    #         vmax=q_breaks[-1]
-    #         # palette=get_palette()
-    #     )
-        
-    #     # Add the raster layer to the map
-    #     raster_layer = RasterLayer(
-    #         url=temp_path,
-    #         bounds=bounds,
-    #         colormap=colormap,
-    #         opacity=input.opacity()
-    #     )
-        
-    #     m.add_layer(raster_layer)
-        
-    #     # Add legend
-    #     for ctrl in m.controls:
-    #         if isinstance(ctrl, LegendControl):
-    #             m.remove_control(ctrl)
-                
-    #     legend = LegendControl({"IWI": colormap}, position="bottomright")
-    #     m.add_control(legend)
-    
+    def _():
+        if input.SelectedMap() == True:
+            m.remove_layer(band_1_json)
+            m.add_layer(geo_json)
+            return m
+        elif input.SelectedMap() == False:
+            for layer in m.layers:
+                if layer == geo_json:
+                    m.remove_layer(layer)
+                    m.add_layer(band_1_json)
+
     # Handle map clicks
     @reactive.effect
     def _():
         # Set up click event handler
-        def handle_map_click(event = None, feature = None, **kwargs):
-            coords = feature['geometry']['coordinates'][0] #extract feature coordinates 
-            latitudes = [coords[x][1] for x in range(len(coords))] 
+        def handle_map_click(event=None, feature=None, **kwargs):
+            # extract feature coordinates
+            coords = feature['geometry']['coordinates'][0]
+            latitudes = [coords[x][1] for x in range(len(coords))]
             longitudes = [coords[y][0] for y in range(len(coords))]
-            country_name= feature['properties']['sovereignt'] #find country name 
-            country_abbrev= feature['properties']['sov_a3'] #find country abbreviation
+            # find country name
+            country_name = feature['properties']['sovereignt']
+            # find country abbreviation
+            country_abbrev = feature['properties']['sov_a3']
 
-            selected_country.set(country_name) #set the country name 
+            selected_country.set(country_name)  # set the country name
 
-            centroid = (np.mean(latitudes),np.mean(longitudes)) #lock view position to the country's centroid 
+            # lock view position to the country's centroid
+            centroid = (np.mean(latitudes), np.mean(longitudes))
             m.center = centroid
             m.zoom = 5
-        
+
         # Register click handler
         geo_json.on_click(handle_map_click)
-    
+
     # Display value boxes
     @output
     @render.text
     def highest_iwi():
         raster_data = selected_raster()
         return f"{np.nanmax(raster_data):.3f}"
-    
+
     @output
     @render.text
     def lowest_iwi():
         raster_data = selected_raster()
         return f"{np.nanmin(raster_data):.3f}"
-    
+
     @output
     @render.text
     def avg_iwi():
         raster_data = selected_raster()
         return f"{np.nanmean(raster_data):.3f}"
-    
+
     # Generate trend plot for mean IWI across Africa
     @output
     @render.plot
     def trend_plot():
-        fig, ax = plt.subplots(figsize=(12, 8))
-        ax.plot(range(len(time_periods)), band_means, marker='o', color="darkorange", linewidth=2, markersize=6)
+        fig, ax = plt.subplots(figsize=(10, 4))
+        ax.plot(range(len(time_periods)), band_means, marker='o',
+                color="darkorange", linewidth=2, markersize=6)
         ax.set_xticks(range(len(time_periods)))
         ax.set_xticklabels(time_periods, rotation=45, ha="right")
         ax.set_ylabel("Mean IWI")
         ax.set_ylim(0.1, 0.3)
         ax.set_title("Average IWI Over Time (Africa)")
         ax.grid(True, linestyle='--', alpha=0.7)
-        
+
         plt.tight_layout()
         return fig
-    
+
     # Generate histogram plot
     @output
     @render.plot
     def iwi_residuals():
         country_name = selected_country.get()
         fig = get_residual_plot(country_name, residual_data)
-        return fig 
-    
+        return fig
+
     # Plot time series at clicked location
     @output
     @render.plot
     def clicked_ts_plot():
         country_name = selected_country.get()
-        
+
         fig, ax = plt.subplots(figsize=(10, 4))
-        
+
         if country_name is None:
             ax.text(0.5, 0.5, "Click on the map to see the IWI time-series here.",
-                   horizontalalignment='center', verticalalignment='center',
-                   transform=ax.transAxes, fontsize=14)
+                    horizontalalignment='center', verticalalignment='center',
+                    transform=ax.transAxes, fontsize=14)
         else:
-            ax.plot(IWI_df['Band_Number'], IWI_df[country_name], marker='o', color="darkorange", linewidth=2, markersize=6)
-            ax.set_xticks(range(1,len(IWI_df['Band_Number'])+1))
+            ax.plot(IWI_df['Band_Number'], IWI_df[country_name],
+                    marker='o', color="darkorange", linewidth=2, markersize=6)
+            ax.set_xticks(range(1, len(IWI_df['Band_Number'])+1))
             ax.set_xticklabels(time_periods, rotation=45)
             ax.set_ylabel("IWI (0 to 1)")
             ax.set_ylim(0, 1)
             ax.set_title(f"Time Series of IWI in {country_name}")
             ax.grid(True, linestyle='--', alpha=0.7)
-        
+
         plt.tight_layout()
         return fig
-    
+
     # Display improvement data table
     @output
     @render.data_frame
@@ -543,15 +554,24 @@ def server(input, output, session):
             filters=True,
             height="800px"
         )
-    
+
     # Download CSV handler
-    @session.download(filename=lambda: f"poverty_improvement_{datetime.now().strftime('%Y-%m-%d')}.csv")
+    @output
+    @render.download(filename=lambda: f"poverty_improvement_{datetime.now().strftime('%Y-%m-%d')}.csv")
     def download_data():
         return improvement_data.to_csv(index=False)
-    
-    
+
+    @output
+    @render.download(filename=lambda: f"{selected_country.get()}_IWI.csv")
+    async def download_country_data():
+        country_name = selected_country.get()
+        buf = io.StringIO()
+        country_data = pd.DataFrame(IWI_df[country_name])
+        country_data.to_csv(buf, index=False)
+        yield buf.getvalue()
+
 
 # ------------------------------
 # 4. Create and run the app
 # ------------------------------
-app = App(app_ui, server)
+app = App(app_ui, server)

data/band_1.geojson

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3622bf0268d2a7141d48495366c4e5c259882c6219cace3d246d336299b567ac
+size 136321755

helpers/reduce_precision.py

@@ -0,0 +1,12 @@
+import json 
+
+def reduce_coordinate_precision(geojson_data, precision=5):
+    """Reduces the decimal precision of coordinates in GeoJSON data."""
+    features = geojson_data.get('features', [])
+    for feature in features:
+        if feature['geometry']['type'] == 'Polygon':
+            coords = feature['geometry']['coordinates']
+            rounded_coords = [[[round(coord, precision) for coord in point] for point in ring] for ring in coords]
+            feature['geometry']['coordinates'] = rounded_coords
+
+    return geojson_data