Houston Crime Explorer

The app allows you to explore the City of Houston crime data set through the use of an interactive choropleth map, line charts, histograms, and an hour-by-hour heatmap. All instances of crime rate refer to the rate per 100,000 persons.

NOTE:

• This app is not mobile ready. Sorry. Pop-ups are sticky; I will get around to fixing this.
• If using Firefox, you will have to select a 'Crime Category' to populate the choropleth map. For whatever reason, the app does not auto-populate the map with the default 'Aggravated Assaults' category.

Tour

The choropleth map gives you a tooltip with some helpful information.

Selecting police beats on the map changes the output of the plots.

Deselecting the police beat resets the plots.

The map changes when the crime type is changed.

You can also change the year.

Build It

The app is available online. However, you can run this app yourself.

To run the app, you will need to:

1. Install R
2. Install RStudio
3. Install Git
4. Clone this project: git clone https://github.com/seasmith/HoustonCrimeViewer HoustonCrimeViewer
5. Install devtools and packrat: install.packages(c("devtools", "packrat"))
6. Install requisite packages: packrat::init() or devtools::install_deps() (the DESCRIPTION file exists for this purpose, this repo is not a package)
7. Execute shiny::runApp()

Please create an issue if something does not go as planned.

Future Versions

Definitive Additions

Some things which will eventually be added:

• A 'data brief' section to document:
  • Data distribution
  • Missing/anomalous data
• plotly maps.
  • Lasso-selection
  • Multi-selection
  • Satellite imagery
• Start- and end-year range selection.
  • Currently, the server gets hung up when the slider is used. May have to resort to using another dropdown menu.
• School district polygons. This may make the map look a little messy, so this may be placed in its own separate map or as a static report per school district.
• Toggle certain beats that skew the crime rates. Right now, this is pre-set (the grayed-out polygons).
• Select which variable to map. Currently rate is used, but others may include:
  • n_offenses (number of offenses)
  • n_offenses / pop_den (number of offenses divided by population density)
• Attach names to the police beats:

Police Beat	Name
1A20	West Montrose
1A30	East Montrose
1A40	River Oaks
1A50	Afton Oaks
2A20	The Heights (East)
2A30	The Heights (Central)
2A40	First Ward
2A50	Rice Military
2A60	The Heights (West)
3B30	Oak Forest
3B40	Independence Heights
4F30	Addicks Reservoir
6B60	Greenspoint
10H40	Midtown
10H50	Third Ward (University of Houston)
10H70	Museum District
12D50	Clear Lake (North)
12D60	Clear Lake (Central)
10H80	Rice Village
14D20	Sunnyside
15E40	The Astrodome
17E10	Gulfton
18F10	Tanglewood (Houston)
18F20	The Galleria
18F30	Fountainview
18F40	Piney Point
19G10	Alief (East)
19G20	Alief (Central)
19G30	Alief (West)
19G40	Alief (South)
20G30	West Chase
20G50	City Centre
20G60	Memorial
20G80	Barker Reservoir

Probable Additions

Some things which may be added:

• Address input. Ideally, a user should be able to enter their address to highlight the police beat in which the live.
• Geocoding all crimes. There are hundreds of thousands of unique block ranges (block ranges are the addressing method used by the Houston Police Department). Focus may be prioritized on recent years, starting with 2017.
• Census data. This may include income, age, race, and other data available through the census.
• Turn this project into a package. Vignettes would act as reports and summaries.
• Vega/Vega-lite graphics.

Other Visuals

I have posted tweets on this data in the past, and likely will continue to do so.

Crime in #Houston from 2010-2017 and Hurricane Harvey's impact #rstats - March 5, 2018

Data sources

Type	Source
Crime data	City of Houston
Police beats	City of Houston
Population data	NASA SEDAC