Data Science CoP - Geospatial Data

This repository is a collection of notebooks to get started with geospatial analysis using open government data. It is companion material for the geospatial data analysis tutorial. The goal is to highlight some of the common idioms with geospatial data.

Contents of the Repository

This is a fairly standard way to organize a project:

• data - folder for datasets included and generated in various steps

• docs - I usually include some background/idea articles to further explain the ideas

• notebooks folder - contains project notebooks

• README.md and environment.yml - explain the flow and how to create the environment

To provide some focus, we'll develop the project, and explore the idioms, using an economic development scenario. Economic development covers multiple topics: 1) jobs; 2) education; 3) businesses; and 4) government policies, to name a few.

For this project we will develop data stories to understand how the State of California and City of Los Angeles support micro, small, and mid-size enterprises. To make things manageable, we'll focus on neighborhood level analysis.

Meaningful analysis always starts with questions. Some of the questions we can answer with open data include:

1. Businesses related:

   • What types of businesses are in Los Angeles?
   • Where are they located?
   • Are there clusters of related businesses?

2. Neighborhood related:

   • Do neighborhoods have similar distribution of businesses?
   • Does a neighborhood have special programs for businesses?
   • Is the neighborhood in a Business Improvement District (BID)?

3. Jurisdiction related:

   • Who are the city council and state representatives for the business?
   • Does the business lie within a Business Improvement District (BID)?
   • Is the business in an opportunity zone?

The notebooks in this project demonstrate common geodata idioms to address these questions:

---

Data Basics:

• 1-data-wrangling.ipynb - Show some basic examples of processing geodata. I've identified three representative datasets:

  • Businesses dataset - csv file from city of la
  • Jurisdictions in LA County - shape file from LA County
  • Opportunity zones from the state - shape file from State of California
  • Note: This folder will create geodata from the city url (no need for lfs)

Spatial Analysis:

• 2-data-fusion.ipynb - We'll look at some simple spatial analysis techniques. Specifically:

  • point in polygon techniques to relate businesses and opportunity zones
  • Use choropleth map to show counts and densities
  • Add new datasets for City Council Districts and Business Improvement Districts (BIDs)

• 3-combined-details.ipynb - Use ipyleaflet mapping techniques for data exploration/aggregation..

  • ipyleaflet-based mapping, overlays, and popups
  • Select one BID (Wilshire) to add businesses to the map
  • NAICS analysis on the business to color code map markers
  • Lot's-o-crap-on-a-map!

Urban Form:

• 4-morphology.ipynb - Techniques to compare neighborhoods

  • Clustering zones
  • Street network measures
    
    

Note on the data used

Since this is a github repo I'm trying to avoid having to use lfs. There are two (bigish) files, the first dataset should be downloaded into the data folder before you start. The second dataset is loaded from data.lacity.org automatically:

1. LA Zoning dataset
2. LA businessess pubished by data.lacity.org

I have included the rest of the data or what is needed is computed from these two.

Notes on the env

Here's a simple recipe to create the env from scratch using anaconda::

1. Make sure Anaconda is installed and git clone this repository from github. cd into the gds-tutorial folder.

2. In the gds-tutorial folder you should see the environment.yml file. Use this to build the env using conda: conda env create -f environment.yml

3. Once the create env is completed, activate the new env: conda activate gds-tutorial

4. At this point you will have a working jupyter lab with the necessary widgets. Launch the lab with juptyer lab.

5. You're ready to explore, understand, develop, ...