Shiny App for violent crime analysis in US
Links |
---|
Deployed Shiny App |
Rendered Milestone Writeup |
Shiny App Source Code |
How have your visualization goals changed?
After some valuable feedback and insights from @Tamara_Munzner and Samuel Hinshaw (@hinshaws) on my initial milestone goals, I decided to re-design my app and changed it's end-user.
Previously the US Crime Data App
was meant for a police commissioner of a particular city and visualization had less context and usability, I have tried to simply for this milestone to get the app up and running first.
The App is meant for any official of Department of Homeland Security
or any Federal criminal jurisdiction
who wants to quickly understand Where are the crimes concentrated? , Which cities have shown increase over decrease in crime rates? between 1995-2015
Apart from overall view of concentration of crimes across the US, the city-wise crime comparison is important to get an insight into how are the cities are performing against each other?
The reason an official will be particularly interested in a city-level comparison is to understand how are the particular jurisdiction performing against each other and which cities need better judicial infrastructure and stronger law and order enforcement
In short, the US Crime Report
app provides a quick-view for concentrated and crime trends in the US based on the year and city (jurisdiction).
gaps
in the app later, also right now the usability is very basic and aesthetics are minimalIntro
provides a summary of app layout, navigation and usability.
So before making any assumptions, the user can understand the layout and what information is being conveyed
Carries information about navigation and how the marshall crime project come About
Additionally carries app developer's info and contact if a user you like to interact or suggest improvements
A high-level overview of what each tab
represent
Map
:
Visualises the number of crimes between 1995-2015 by selecting the crime type based on the geographic concentration of crimes. The added feature of Show crime relative to Population
normalizes the crime numbers by the population of the city
The instructions about zooming and viewing data labels using tooltips are included
Since it is a leaflet map, and not the most sophisticated of using map tools the user can find little problem initially to adjust the view and aspect relationship
WIP features: Instead of showing raw crime numbers for a city and all in green
I will be implementing a colour scale
of increase or decrease in crime since past 5 years, the colour scale will be red, yellow, green on a gradient scale based on increase in decrease in crimes
Trends
: Shows the trend of all crimes between two cities based on the user selection of time period and cities
prediction
or forecast
for years after 2015, that is 2016-2018Data
: Based on year and city selection user can download data as a csv
Who made it? This app was created by Akshi Chaudhary, masters student of data science at UBC
If you'd like to get in touch; talk about this app, contribute to it, or make a suggestion for new features, contact via twitter @ChaudharyAkshi or on github akshi8