Charlotte in a Nutshell (10 Maps)
Fall is upon us and the school year is starting to catch up. This fall I'm taking several interesting classes including Machine Learning, Geospatial Analytics and Network Science. For one of my classes, I have a class project in which I'm going to explore crime patterns for Charlotte (my hometown -- yes, there are some people originally from Charlotte). The data set is from the CMPD and covers all crimes (and reported incidents) from 2012-2014.
However, as with any data science/analytics project, much of the initial work is on data preprocessing/wrangling. Fortunately, I have a great teammate who's focusing more on the data steps while I get to do what I do best -- explore data! The goal of my exploration is to come up with good questions (hypotheses) that I want my analysis/modeling to answer.
My initial hypothesis is that location matters in (Charlotte) crime... in data speak, there is a "spatial" aspect to the data. It may seem straightforward, but how to model location (spatial) is not easy and is itself its own field in statistics and data mining.
To help my exploratory analysis, I've used arcGIS Online (www.arcgis.com) to create ten initial plots of Charlotte across several demographic. The data comes from data provided through arcGIS (http://www.arcgis.com/home/item.html?id=36d83cc2db854958a092254999992e8f)
What's interesting to me is how clear it is that Charlotte is "segregated", i.e. from these maps show that there are very differentiated areas of Charlotte by demographic factors. Can you see them?
(One reference I'm going to read more up is a book by Dr. Tom Hanchett of the Levine Museum of the New South, "Sorting Out the New South City: Race, Class, and Urban Development in Charlotte, 1875-1975", see http://www.historysouth.org/charlottebooks)
Please leave any thoughts, comments or ideas (hypotheses) to help me... thanks!
1. Number of Violent Crimes Divided by Population
2. % of Population that is Black
3. % of Population that is White
4. Single Family Home Median Property Value
5. Median Income
6. Percent of Population without a High School Diploma
7. Percent of Population on Food Stamps
8. Percent of Population on Medicaid
9. Voter Participation
10. Births to Adolescent Population
Great maps Ryan W.. Charlotte is part of the national open data project for police data which is a great endeavor for us as a city to participate in. As you continue to discover, what about overlapping some data that will help get to the "what can we do about it?" question. Does the CMPD source (or other open data) also contain information around police patrols, charitable distribution centers, distance to polls, or distance to healthcare providers? This is the type of work that Charlotte can leverage to become an even better city!
Love where this is going - great work so far!