focusing on the interaction aspects and connecting these choices to a data abstraction (including a characterization of the raw data types and their scale/cardinality, and of any derived data that you decide compute)
I have utilized the following design to effectively do the job of comparing the criminal records and trends across different cities. Shown below is the dashboard of the same.
The slider input takes in a range of the year and then uses that information to plot it on the graph.
The paramter radio button takes in a parameter value from a range of values and then uses them on the plot's y-axis. The paramters can range from homicides, rapes, violent crimes and assaults. I wanted this to be a radio button because there can be only one parameter that can be plotted on the y-axis.
The select input button can be used to select the city to compare with others. This will then be used to filter out the dataset to include just that city.
The Download results button can be used by the user to download the data table as a CSV onto their own computer.
The graph displays the trend in crime data as compared to all the other cities in the Marshall dataset.
You should also concisely describe your visual encoding choices.
I wanted to see the trend in a particular city's crime rate as opposed to other cities. This could be done by highlighting the trend in red and plotting the rest of the trend lines with low opacity.
I had initially set out to design an app that would display the translink data on a shiny app. But I ran into some troubles with extracting the latest bus locations so I had to stall this project for the moment, and have instead developed this app where I am analyzing the Marshalls Crime data. The first version of this app does do all the functionality I have descibed in my proposal. I aim to build on this further by adding some more graphs and analysis.