travishorn / poi-classifier

Geospatial Point of Interest (POI) Classifier

Objective

A geospatial point of interest (POI) classifier that takes in data points with latitude, longitude, and name information and predicts the category of each point of interest. This project demonstrates data preprocessing, feature engineering, and classification using geospatial data.

Supported categories are: National Battlefield Site, National Battlefield, National Heritage Site, National Historic Park, National Historic Reserve, National Historic Trail, National Lakeshore, National Memorial, National Military Park, National Monument, National Park, National Preserve, National Recreation Area, National Recreational River, National River, National Scenic River, National Scenic Trail, National Seashore, Other, Park, and Parkway.

Data Source

National Parks, Monuments, Preserves, & Historical Parks from POI Factory: http://www.poi-factory.com/node/31147

Methods Applied

1. Data Collection: I gathered geospatial data points for different POIs. Each data point had latitude, longitude, and a name of the POI (e.g., Fort Dupont Park - 3600 F St SE Washington DC 20019>202-426-7723).
2. Data Preprocessing: I processed the data, handling missing values, removing duplicates, and performing any necessary data transformations. I added type information to categorize each point (e.g., National Park, National Heritage Site, etc.) for 80% of the data, reserving 20% for demonstration purposes.
3. Feature Engineering: I extracted relevant features from the geospatial data that could be useful for classification. In this case: latitude, longitude, and name, as they were the only data points available in the data.
4. Model Selection: I chose and implemented a random forest classifier using the popular machine learning library scikit-learn for Python.
5. Model Training: I split the data into training and testing sets and trained your classifier on the labeled data.
6. Model Evaluation: I evaluated the classifier's performance using appropriate metrics such as accuracy, precision, recall, and F1-score.
7. Real-world Usage: I used the trained model to classify the remaining 20% of data. The result is a complete set of national parks, monuments, preserves, and historical parks in the United States, classified with their type.
8. Data Visualization: I created a map of all POIs using Folium which colors each marker differently depending on the type. This includes the types from the predicted set.

Prerequisites

• Python
• scikit-learn, pandas, and folium installed through pip
• If you want to use the pre-trained model on your own data: a CSV file containing latitude, longitude, and names of POIs
• If you want to train your own model: a CSV file containing latitude, longitude, name, and types of POIs
• If you want to use the same source data as I did: a copy of the data from POI Factory

Usage

Clone this repository

git clone https://github.com/travishorn/poi-classifier

Change into the directory

cd poi-classifier

Train

If you want to train your own data, make sure the training set is a CSV file named poi_training.csv with columns for latitude, longitude, name, and type.

Train the model

python train.py

The script will reserve 20% of the training file for testing itself. Scores are shown and trained_model.pkl is created (or overwritten).

Classify

The trained model saved at trained_model.pkl can be used to classify types of POIs. Make sure the "to classify" set is a CSV file named classify.csv with columns for latitude, longitude, and name.

Predict (classify) the POIs

python classify.py

A sample of the predicions are shown and poi_classified.csv is created which should mirror poi_to_classify.csv, but with a type column.

Visualize

Generate a map of POIs using a combination of poi_training.csv and poi_classified.csv, which should now include a complete set of national parks, monuments, preserves, and historical parks in the United States. Each marker is colored according to the POI's type (e.g., national scenic rivers are dark blue, national memorials are beige, etc.).

License

The MIT License

Copyright 2023 Travis Horn

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the “Software”), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.