🌲 HeatMappers Forest Modeling Project

This repository contains our team's work across Sprints 1–4 for the forest modeling and classification challenge. The goal of this project is to develop machine learning models that predict key forest attributes — such as DBH (Diameter at Breast Height), CBH (Crown Base Height), species, genus, and plant functional type (PFT) — using a combination of Aerial LiDAR (ALS), Terrestrial LiDAR (TLS), and imputed treelists from FastFuels.

---

📊 Datasets Used

External Datasets

• FIA Database: CA_TREE.csv, CA_PLOT.csv, REF_SPECIES.csv
• Field Survey Data: 03_tree.csv, 01_plot_identification.csv
• ALS/TLS LiDAR Data: USGS .laz files and .ptx TLS files
• FastFuels TreeMap 2016: Treelists generated using the FastFuels API
• Geo Boundaries: independence_lake_boundary.geojson, sedgwick_boundary.geojson

Derived Datasets

• ttops.csv: ALS-derived tree tops detected from CHM and peak detection
• tls_treelist.csv: TLS-derived tree-level data
• fftl_plots.csv: Predicted treelist from FastFuels for target plots
• combined_intelimon_metrics.csv: TLS-derived structural metrics from IntELiMon

---

🔧 Required Libraries

To run the notebooks, make sure you have the following Python libraries installed:

• numpy
• pandas
• matplotlib
• seaborn
• scikit-learn
• xgboost
• imblearn
• geopandas
• rasterio
• laspy
• pyproj
• folium
• plotly
• pdal
• skimage

---

▶️ Running Instructions

Follow these steps to run the project successfully:

1. Clone the repository
   Open a terminal and run:
   https://github.com/Anshu-b/HeatMappers

2. Install dependencies (different for each sprint)

   pip install -r requirements.txt
   

3. Start Jupyter
   Run either:
   jupyter notebook
   or
   jupyter lab

4. Open and run the notebooks

   • Notebooks are organized by sprint (see below).
   • Run all cells from top to bottom.
   • Outputs include classification reports, graphs, CHMs, and CSV files.

---

🗂️ Notebooks with our code by Sprint

Sprint 1

In order to run sprint 1a make sure unzip data.zip

• sprint1_1-4_heatmappers.ipynb
• sprint1_5-6_HeatMappers.ipynb

Sprint 2

• sprint2_1-2_HeatMappers.ipynb
• sprint2_3-4_HeatMappers.ipynb

Sprint 3

In order to run sprint 3 make sure unzip data.zip

• sprint3_HeatMappers.ipynb

Sprint 4

In order to run sprint 4 make sure unzip data.zip

• sprint4-1-field_HeatMappers.ipynb – PFT, genus, and species classification from field data
• sprint4-2-tls-enhancement_HeatMappers.ipynb – Predicting CBH and CR
• sprint4-3-ff_HeatMappers.ipynb – Classification using FastFuels + predicted features

---

📦 Expected Outputs

• Visualizations: CHMs, confusion matrices, histograms, treelist charts
• Model metrics: Accuracy, F1-score, precision, recall
• Classification reports per model/task
• CSV outputs like ttops.csv, fftl_plots.csv
• Final report summarizing all results and findings

---

📄 Final Report

Please see HeatMappers_project_report.pdf for:

• Model comparison (Random Forest, Logistic Regression, AdaBoost)
• PFT, genus, and species prediction performance
• Accuracy analysis of field-based vs. FastFuels-based models
• Discussion

---

👥 Team

Team Name: HeatMappers
Members:

• Ansh Bhatnagar
• Jiahe Qin
• Jeronimo Adames-Baena
• Ishayu Ghosh

---

📬 Submission Notes

• All notebooks for Sprints 1–4 are included in this repository.
• Final report is submitted as: HeatMappers_project_report.pdf
• URL to this repository is included in the Sprint 4 text submission.