Visualization with Tableau

Making Data Understandable

This week, in HCDE 210 studio, the focus was on creating visualizations for a set of data. Visualizations are used to easily compare data, keep data organized, and display data clearly. In order to help humans perceive data easier, visualizations include different colors, densities, proportions, textures, and/or orientations. To make our own visualizations, we downloaded a program called Tableau, which allows users to create visual representations of large sets of data. We did a pre-studio tutorial with data on Seattle bike racks, which was helpful. I liked getting familiar with Tableau and figuring out the features before doing the sprint. At first, I struggled to create visuals other than bar graphs and line graphs, because that’s what I’ve mostly made in the past, but I now know how to make different types of maps and charts, as well. I also struggled with filtering out certain sets of data in the beginning, but after practicing, I was more comfortable with it, and was able to create the visuals for the sprint without too much trouble.

Seattle 911 Calls

We used Seattle 911 call data for this week’s sprint. The scenario I came up with to answer with visualizations is: “a woman is planning to open a business in the U-District and wants to know what areas experience the least theft and robbery. Where should she open her business, and during which business hours should she be most prepared for an act of theft or a robbery?”.

This is a tree map representing the most dangerous streets for the woman to open her business in terms of recorded shoplifting, armed robbery, and commercial burglaries. The dark blue square in the top left (43XX University Way NE) is the block with the most recorded incidents, while the square in the bottom right represents the least recorded incidents.
This bar graph shows how many theft incidents have been recorded at particular business hours on specific streets in U-District. For example, hour 16 on 43XX block of University Way NE is the most dangerous at 54 recorded theft incidents, while on 45XX block of University Way NE, the most dangerous hour is 15.

I then created three visualizations using the 911 call data, keeping Tufte’s Visualization Principles (avoid distorting data, encourage the eye to compare, show data at several different levels of detail, and make large data sets coherent), to make the most dangerous streets and most dangerous times apparent. In one visualization, I created a map comparing coordinate locations in U-District with the number of recorded theft incidents at that location. In this visual, the user can compare the size of dots that represent the number of reported theft incidents in certain areas. I also made a tree map of incidents recorded on streets in the U-District, so that it is easy to see which streets are the most dangerous. The tree map allows the user to compare the number of incidents visually, as the more incidents that are recorded, the larger and darker the block that represents it. The third visualization I created is a bar graph that shows how many theft incidents have been recorded during certain business hours on streets in the U-District.

These visuals are published on my Tableau Public profile and can be viewed at https://public.tableau.com/profile/publish/911calls_2/BetterU-DistrictTheft#!/publish-confirm

Reflection:

This project opened my eyes to a whole other part of HCDE. So far, in this class, we have focused a lot on determining issues with products and observing the way humans interact with certain objects, to devise a better way for things to be constructed. Now, I see that HCDE is not just about collecting data, but also presenting it in a clear and organized way, so that it is useful and understandable. During the project I wondered what the visuals would look like if the data set included the Seattle 911 calls for the whole year, not just a month. Would the differences among the bubbles in the packed bubble visualization be more or less extreme? I think it would be interesting to see how the amount of data used to make a visualization effects how it appears.

The Inquiry:

How Would the Release of These Visuals Representing U-District Theft Incidents Affect the Status of Certain Areas of the District?

It is that visualizations can have unintended consequences. For example, possible that if these visualizations on theft in U-District were released to the public, new business owners could access them when trying to pick the best location to open their store, and the most dangerous streets might experience a lack of new businesses. I think that this information could create a greater divide in the city in terms of safety, but also in terms of social status. It is possible that streets that experience less theft would cost more to open a business on than those that experience a lot of theft. This means that more wealthy people would be able to afford safer locations, while people of lower socioeconomic status would only be able to afford the more targeted locations.