Welcome to GIS Hell (It’s Actually Pretty Cool)

GIS = Geographic Information Systems = “I put data on maps and somehow this became my job”

What you’ll learn: - How to not mess up coordinate systems (spoiler: you will anyway) - Make pretty maps that impress your boss - Pretend you know what you’re doing

Install This Stuff (Or Cry Later)

# R packages (install once, debug forever)
install.packages(c("sf", "leaflet", "tmap"))

# Python (because why make life easy?)
# pip install geopandas folium

Chapter 1: Your First Map (That Won’t Suck)

The Magic of Putting Dots on Maps

library(sf)
library(leaflet)

# Create some fake data (like your confidence)
places <- data.frame(
  location = c("Coffee Shop", "My House", "Regret"),
  lon = c(-122.4, -122.5, -122.6),
  lat = c(37.8, 37.9, 37.7)
)

# Turn it into spatial data (✨ magic ✨)
spatial_places <- st_as_sf(places, 
                          coords = c("lon", "lat"), 
                          crs = 4326)  # 4326 = "GPS coordinates"

import geopandas as gpd
from shapely.geometry import Point

# Same fake data, different language
data = {
    'location': ['Coffee Shop', 'My House', 'Regret'],
    'lon': [-122.4, -122.5, -122.6],
    'lat': [37.8, 37.9, 37.7]
}

# GeoPandas: Pandas but with existential crisis about location
gdf = gpd.GeoDataFrame(data, 
                      geometry=[Point(x, y) for x, y in zip(data['lon'], data['lat'])],
                      crs='EPSG:4326')

Make It Interactive (Impress People)

# Leaflet: Like Google Maps but you made it
leaflet(spatial_places) %>%
  addTiles() %>%  # Free maps (thanks OpenStreetMap!)
  addMarkers(popup = ~location) %>%
  setView(lng = -122.5, lat = 37.8, zoom = 10)

import folium

# Folium: Python's answer to "I want pretty maps"
m = folium.Map(location=[37.8, -122.5], zoom_start=10)

for idx, row in gdf.iterrows():
    folium.Marker([row.geometry.y, row.geometry.x], 
                 popup=row['location']).add_to(m)

m  # Boom. Map.

Chapter 2: Buffers (Circles Around Stuff)

Why buffers? “Everything within 1 mile of Starbucks” is a legitimate analysis.

# Transform to meters (because measuring in degrees is weird)
places_meters <- st_transform(spatial_places, crs = 3857)

# Make 500m circles around each point
buffers <- st_buffer(places_meters, dist = 500)

# Back to normal coordinates for mapping
buffers <- st_transform(buffers, crs = 4326)

# Add to map
leaflet() %>%
  addTiles() %>%
  addPolygons(data = buffers, color = "blue", fillOpacity = 0.3) %>%
  addMarkers(data = spatial_places, popup = ~location)

# Same thing, more typing
gdf_meters = gdf.to_crs('EPSG:3857')
buffers = gdf_meters.buffer(500)  # 500 meters
buffer_gdf = gpd.GeoDataFrame(gdf[['location']], 
                             geometry=buffers).to_crs('EPSG:4326')

# Add to map
m = folium.Map(location=[37.8, -122.5], zoom_start=10)

# Add buffers
for idx, row in buffer_gdf.iterrows():
    folium.GeoJson(row.geometry, 
                  style_function=lambda x: {'color': 'blue', 'fillOpacity': 0.3}).add_to(m)

# Add points
for idx, row in gdf.iterrows():
    folium.Marker([row.geometry.y, row.geometry.x], 
                 popup=row['location']).add_to(m)
m

Chapter 3: Distance (How Far is Too Far?)

# Calculate distances between all points
distances <- st_distance(spatial_places)
print("Distances in meters (approximately):")
print(distances)

# Convert to miles because America
distances_miles <- distances * 0.000621371
print("Distances in freedom units:")
print(distances_miles)

from sklearn.metrics.pairwise import haversine_distances
import numpy as np

# Extract coordinates and calculate distances
coords = np.radians([[p.y, p.x] for p in gdf.geometry])
distances = haversine_distances(coords) * 6371000  # Earth radius

print("Distances in meters:")
print(distances.astype(int))

Chapter 4: The Coordinate System Nightmare

Truth: You will spend 50% of your GIS career fixing coordinate systems.

Common CRS (Coordinate Reference Systems): - EPSG:4326 = GPS coordinates (lat/lon) - Use for web maps - EPSG:3857 = Web Mercator - Use for measurements - Local UTM zones - Use when you need to be precise

# Check what CRS you're using
st_crs(spatial_places)

# Transform between coordinate systems
utm_data <- st_transform(spatial_places, crs = 32610)  # UTM Zone 10N
print("Now in UTM coordinates:")
print(st_coordinates(utm_data))

Chapter 5: Quick Wins

Load Real Data (Finally!)

# Load a shapefile (if you have one)
# data <- st_read("your_file.shp")

# Or get data from the internet
library(osmdata)
# Get coffee shops in your area
# coffee <- opq("San Francisco, CA") %>%
#   add_osm_feature(key = "amenity", value = "cafe") %>%
#   osmdata_sf()

Export Your Masterpiece

# Save as shapefile
# st_write(spatial_places, "my_awesome_data.shp")

# Save as GeoJSON (the cool kids format)
# st_write(spatial_places, "my_awesome_data.geojson")

The End (You Survived!)

What you learned: - Points on maps = happiness - Buffers = circles around things - CRS = the source of all problems - GIS = surprisingly addictive

Next steps: - Make more maps - Break more coordinate systems
- Blame ESRI for everything - Profit???

Pro tip: When in doubt, reproject to EPSG:4326. When that doesn’t work, try EPSG:3857. When that fails, panic quietly.

“In GIS, all roads lead to Rome… if your coordinate system is correct.”