Team Members: Dennis Orzikh, Alexander James Oscar Craver Kirchhoff, Bryant Wong, Joseph Cutrell

Design Rationale

We felt that since the data contained geographic information, we wanted to create a map with filters the user could explore. We chose to use color instead of area or other encodings for the data because we wanted to show trends and also make them clear to the average Seattle resident. The plan was to build on this prototype, allowing users to drilldown if they want more specific data. We feel that using color works well, as it allows the user to quickly eyeball trends in our data. We decided on green since it is a natural map color for land. The particular filters were chosen by the columns available to us. We designed the legend in order to show both the distribution of incidents as well as communicate the color scheme effectively. However, after creating this legend, we now realize that there might be better ways to encode the distribution (histogram maybe) We made the choice to allow the visualization to normalize with each filter change, since we felt that this was most effective in communicating how beats compare. Besides checkboxes for filters, we provide the option for the user to hover over both beats on the map and legend.

Development Process

We came up with a list of features that would be feasible to complete as a prototype, then implemented them from high level to fine grained. We started by just making a beat map, then adding color/data, and finally interactivity via filters and hover, and animations. We added the legend at the end, first with a static legend that only showed color information, then made it interactive and rescale, since this added more visual weight to the colors. The vast majority of the prototype was done with everyone working together on one computer, with certain features done individually. We estimate it took about 64 people hours to develop (3 sessions of 5ish hours with 4 people). The hardest part was actually figuring out how to use reference systems of shapefiles to turn them to GeoJSON, as well as drawing the map in D3 using map projections.