Upload 5 files

README.md

@@ -0,0 +1,21 @@
+---
+title: GeoMetals App
+emoji: 🌍
+colorFrom: blue
+colorTo: green
+sdk: streamlit
+sdk_version: "1.30.0"
+app_file: app.py
+pinned: false
+---
+
+# GeoMetals App
+
+This app uses a trained Gradient Boosting regression model to estimate heavy metal contamination in soil based on geolocation.
+
+### Features:
+- Interactive map for geospatial input
+- Predicts Fe, Cr, Mn, Mo, In, Ta in ppm
+- Warns if values exceed EPA thresholds
+
+Trained on SEM-EDS environmental sampling data from Arizona.

app.py

@@ -0,0 +1,33 @@
+# app.py
+import streamlit as st
+import folium
+from streamlit_folium import st_folium
+from model import predict_metals
+
+st.set_page_config(page_title="GeoMetals Predictor", layout="wide")
+
+st.title("Heavy Metal Concentration Estimator 🌎")
+st.markdown("Click on the map to select a location. The app will predict heavy metal concentrations based on geolocation and trained SEM-EDS data.")
+
+with st.expander("How it works"):
+    st.markdown("""
+    This app predicts concentrations of heavy metals in ppm (parts per million) using a trained Gradient Boosting model. 
+    It accounts for location-based proximity to industrial zones and was trained on SEM-EDS data from Arizona.
+    """)
+
+default_coords = [33.4484, -112.0740]  # Phoenix center
+
+m = folium.Map(location=default_coords, zoom_start=10)
+click = st_folium(m, height=500, width=700)
+
+if click and click["last_clicked"]:
+    lat, lon = click["last_clicked"]["lat"], click["last_clicked"]["lng"]
+    st.success(f"Selected Location: Latitude {lat:.4f}, Longitude {lon:.4f}")
+
+    prediction = predict_metals(lat, lon)
+    st.subheader("Predicted Heavy Metal Concentrations (ppm):")
+    for metal, val in prediction.items():
+        warning = ""
+        if (metal == "Fe_ppm" and val > 300) or            (metal == "Cr_ppm" and val > 100) or            (metal == "Mn_ppm" and val > 200) or            (metal == "Mo_ppm" and val > 10) or            (metal == "In_ppm" and val > 1) or            (metal == "Ta_ppm" and val > 0.5):
+            warning = "⚠️ Above EPA Threshold"
+        st.markdown(f"**{metal}**: {val:.2f} ppm {warning}")

baseline_heavy_metals_multi.csv

@@ -0,0 +1,10 @@
+lat,lon,Fe,Cr,Mn,Mo,In,Ta
+33.3,-112.3,27415.86769483495,51.3699669529441,373.942541888569,1.5349232940362272,0.053601690021355154,0.006718599374013687
+33.3,-112.1,26299.1057740642,44.48211950049188,406.6022298641086,0.9348349532298319,0.059686284846273446,0.005645732788723248
+33.3,-111.9,26818.234649837654,44.2797535086274,403.8412320567538,1.3555984952160824,0.03851945093844259,0.004397743797117824
+33.5,-112.3,22271.049325705106,49.57302757797042,334.7535249204071,2.109642757787766,0.03235873100699918,0.006499888675675197
+33.5,-112.1,23637.91893464434,47.805165337355305,377.42659899019395,2.002124447817415,0.05043702668084274,0.006073682012098753
+33.5,-111.9,25114.072059844577,46.76479987769451,418.4831996205925,1.3413828043500886,0.037197088717647445,0.003698396674258
+33.7,-112.3,25076.263853704233,42.5567995730596,335.6242972045018,1.6538823752374487,0.03781510358713415,0.003839674974730694
+33.7,-112.1,24937.560948569615,40.49942170963043,403.67516495018555,2.0920382674807674,0.046209423538176794,0.003940420853783351
+33.7,-111.9,25272.260466440966,41.12384103166156,400.7125444599324,1.5946907930043717,0.04964803061537517,0.005056558949589168

geo_metals_model.joblib

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03bf0dd7e5f6c26b24bebd200665f5f798c1b6054104623287c7fcfbf6216fcf
+size 594126

model.py

@@ -0,0 +1,15 @@
+# model.py
+import joblib
+import pandas as pd
+
+MODEL_PATH = "geo_metals_model.joblib"
+model = joblib.load(MODEL_PATH)
+
+def predict_metals(lat, lon):
+    input_data = pd.DataFrame([{
+        'Latitude': lat,
+        'Longitude': lon
+    }])
+    prediction = model.predict(input_data)[0]
+    metals = ['Fe_ppm', 'Cr_ppm', 'Mn_ppm', 'Mo_ppm', 'In_ppm', 'Ta_ppm']
+    return dict(zip(metals, prediction))