Fork me on GitHub

Project Team 9: Pedestrians, DS 4200 F19

Kyle McCrosson, John Clancy, Philip Wesley

Service-Learning Course Project as part of DS 4200: Information Presentation and Visualization, taught by Prof. Cody Dunne, Data Visualization @ Khoury, Northeastern University.

Motivation

Chester Park is an amazing resource for the residents of the neighborhood. However, the bisection of the park by Massachusetts Avenue makes it difficult for people to utilize both halves, effectively cutting residents' green space in two. Additionally, the lack of any type of pedestrian crossing across Massachusetts Avenue between Tremont Street and Shawmut Avenue creates an unsafe environment where pedestrians are jaywalking across Massachusetts Avenue out of convenience. We counted more than 30 jaywalkers in only one hour across Massachusetts Avenue, in order to avoid the 4-minute walk that is required to cross at either Tremont or Shawmut. We aim to provide an exploration of pedestrian behavior and crossing data in order to find potential solutions to make it more practical to cross Massachusetts Avenue safely, and to give the residents of Chester Square easy access to the full extent of the park. By balancing the advantages and disadvantages of installing some type of pedestrian crossing, we hope to provide the Chester Square Neighborhood Association with a tool to lead the pedestrian crossing discussion.

Our visualization focuses on the following options for a pedestrian crossing:

High-Visibility Crosswalk
High-Visibility Crosswalk

Rapid Flashing Signal
Rapid Flashing Signal Crossing

Pedestrian Hybrid Beacon
Pedestrian Hybrid Beacon

Visualization

Demo Video

Visualization explanation

Our visualization is based on three main views that highlight the major pro's and con's of each choice of pedestrian crossing. The options we explore include leaving the road as is (pedestrians either cross at Tremont/Shawmut, or jaywalk), adding a crosswalk in the middle of the block, adding a crosswalk with a rapid flashing pedestrian sign, and building a Pedestrian Hybrid Beacon which is essentially a stoplight over a crosswalk that would give pedestrians the right of way. For simplicity, we refer to the Tremont/Shawmut option only as Tremont because the travel times, costs, and safety of each of those routes are the same, and discussing them separately is redundant.

In the lower left corner we have a map of the neighborhood, with the five options for pedestrian crossing laid out over it. This provides the user with a spatial awareness of the various intersections involved, as well as the context of where the proposed pedestrian crossing options would be installed. Hovering over each path with the mouse will provide a tooltip label to tell the user which route they are looking at, and will highlight the encodings related to that route in the other graphs on the right side of the visualization.

The bottom right corner contains a bar graph that compares the average amount of time that it takes for a pedestrian to walk the corresponding route on the map. The amount of time is encoded with length on an aligned axis and the route that the bar is referring to is encoded with color. Comparing the length of the bars on an aligned axis is the most effective way to relate that information to the user, and using the same colors to encode the bars as we did the paths over the map (along with linked highlighting and tooltip labels) unites the information from the map with the information from this bar graph.

In the top right corner we have a violin plot that provides information on the distribution of costs of various solutions we felt could effectively connect the parks in a more effective manner than currently is available. The distribution is encoded within the width of the plot at a given point, and the color encodes which route's distribution you are looking at. Using the width it is easy to compare which have distributions that skew higher or lower because they are using the same y scale so it's just a matter of comparing various widths and the height on the y axis of those widths. We liked that this made it easy to get all of the data in perspective unlike a bar graph of just the average or something along those lines.

In the top left corner we have a key which will highlight whichever route you hover over to make it easy to compare across our various graphs in the visualization, and we also ahve a filter box that allows you to narrow down the scope and more easily compare a smaller number of routes. This is especially helpful for the average walking time, because Tremont raises the number of seconds needed for the graph dramatically which makes it harder to see the differences between all the other options.

Data Analysis

Our data collection process involved gathering the spatial locations of all the routes we wanted to analyze, recording the amount of time it takes to traverse each of the routes, and locating cost figures for each. To gather the travel times, we recorded ourselves and other participants walking each of the five routes in our analysis. Since the proposed options (crosswalk, flashing signal, and PHB) don't yet exist in the Chester Square neighborhood, we had to find implementations of them surrounding intersections of Massachusetts Ave in order to get accurate times. The cost data that we gathered came from a 2016 report by the UNC Highway Safety Research Center. Since we were not dealing with huge or complex data sets, there was not much data processing to be done. More of the power in our data comes from the storytelling around the layout of Chester Park and how it might be improved.

According to the travel time data, the amount of time it takes to cross from one side of the park to the other without jaywalking is almost 11 times greater than crossing the street directly (based on avg_travel_times.csv). Jaywalking, crosswalks, rapid flashing signals, and pedestrian hybrid beacons all take a comparable amount of time to cross. Although from this perspective jaywalking seems like a relatively inviting option (it is quick and already is in place) it should be noted how much more dangerous jaywalking is as compared to any other option of route. The outlier of this dataset is the legal walking route-- walking between the halves of Chester Park by crossing at Tremont or Shawmut is by far the least efficient.

Because Massachusetts Avenue is a busy street, it would make the most sense to go for a pedestrian hybrid beacon. This method of crossing completely stops traffic while pedestrians traverse the crosswalk. While the average walking time is slightly greater compared to a plain crosswalk, it would add to safety. The downside is that this is also the most expensive option, costing about 3.5 times as much to build. However, the higher cost of this crosswalk would be offset by the additional safety, where a single potential accident on this road could wipe out all the savings by not placing a signal. but this argument could be used to argue that any sort of crosswalk would save money overall, because of time saved by those who would have gone around, and injuries by those who are jaywalking. In general, the boosts in safety and speed are countered to some degree by an increase in cost as well, so the best solution isn't completely cut and dry.

Task Analysis

Domain TaskHigh-LevelMid-LevelLow-Level
Compare different options of park connectors and evaluate the overall effects of each Compare Explore Discover
Find the potential downsides of each park connector solutions. This will be important for us to research further. Identify Explore Discover
Examine the number of pedestrians that cross Mass Ave at Tremont and Shawmut, as compared to jaywalkers. Compare Explore Discover
Determine which park connector solutions are viable based on their downsides. This will be important to the Chester Square Neighborhood in making a decision. Identify Lookup Discover

The largest overarching goal for our visualization was to provide an exploratory tool to facilitate discussion within the Chester Park Neighborhood Association about how to best address the problem of a pedestrian crossing in Chester Park. We wished to have an interactive map along with graphs of travel time and cost that would help tell the story of the pro's and con's of how best to connect the parks. The map could be used to demonstrate distance and give a spacial representation of the walking distance between the parks and walking distance to get between them. The reason we wanted to do this was to explore our important domain task of determining which solutions are the most and least viable from different perspectives.

Our intention is that once this exploratory tool is created, it can be a source of many constructive conversations around improvements of Chester Park. Our research not only focuses on convenience for area residents, but on the safety of the pedestrians in the area as well. The data we seeked to gather focuses on how many people are jaywalking across Massachusetts Avenue, a very dangerous behavior that puts any person at risk. The urgency of this data will hopefully spur action among the neighborhood association.

Initially, one of our tasks was to look more closely at pedestrian traffic. While we gathered data on pedestrian behavior at the Tremont and Shawmut intersections as well as jaywalking behavior, none of that has so far been included into our visualization because we could not find a meaningful way to include that information without cluttering our view. However, this information is still important for an overall conceptualization of this issue.



Design Process

The first thing we worked on was setting a goal for what we wanted to do. This included planning out what data we intended to collect for our project. We then collected data over the course of several days based on the plans we had made by standing at the intersections, and counting how many people were crossing and in what manner. We also timed ourselves and others to see how long it takes to get to one side of the park to the other.

Once we gathered the data, we moved on to create prototypes that would utilize our data. We did not show all the data in our visualization, as some of that data wasn't going to fit in well with the rest, but gathering an abundance of data was useful in having a good selection to choose from when in came time to creating the visualization. Our best first sketches all shared one common feature: a map of Chester Square with line representations of the routes we wanted to discuss. Spatial location is the most accurate way to reflect data and there is no better way to set the context of this project. From there, we knew that we wanted a collection of accompanying charts to display some sort of useful information about these walking routes.
First Draft Viz Sketch

The sketch prototypes were useful for gauging what types of visualizations worked and which ones did not. from this, we moved on to create a digital prototype of the sketch using PowerPoint. PowerPoint proved to be useful sketch tool that was able to express a degree of interaction using slides with demo interactions. The feedback that we received on the digital sketch revealed that we had tried to shove too much information into a small frame, creating an overload for the users. This was when we decided to remove the safety aspect from our drafts, as we were struggling to come up with a mathematical way to represent and compare safety anyway.
Second Draft Viz Sketch

After that point you can see our final draft of the visualization emerge as it is above. Color is used to encode the walking route of each piece of data across the whole visualization to give everything a unified feel. In the map, routes are mapped out spatially with a line. In the bar chart, travel time is encoded with length on an aligned axis so that comparison is simple. For the violin chart, cost is encoded with length on an aligned axis as well, and distribution proportion is encoded with width of the violin shape. Hovering over any piece of data in these three charts will highlight the data corresponding to that route in the other two charts, as well as display a tooltip with more specific information relevant to that chart. Additionally, the box in the upper left hand corner will allow you to filter the information that is displayed in all three charts so that you do not have to look at all of the routes at once. Doing this will adjust the axes of the bar and violin charts so that you can get a more accurate view of the information being displayed.

Conclusion

Over the course of this project we have spent a lot of time planning our data collection, implementing that data collection, and creating a viable and useful visualization off of that material. The primary focus has been on the difficulty and danger involved in using both sides of the park - there is not effective way to cross from one side of the park to the other. Our main takeaway is that there is no perfect solution to this issue, rather there is an array of possibilities for the neighborhood association to discuss. We hope that the data we have collected and presented aids their discussion and deliberation of the matter. If we could go back and improve, we would try and collect data from more robust sources - what we ended up with provides interesting results, but doesn't allow for as complex of a visualization as we would have liked. This would make for a more engaging outcome.

Going forward, we would like to shift more towards the aspect of safety. We tried to provide some focus to the matter but it did not end up making the final cut of our visualization. Doing enough research and data collection to accurately and effectively portray the relative safety of each potential pedestrian crossing would allow us to create a more comprehensive view of the crossing issue. This would provide another important factor in comparing different solutions. Additionally, it would make sense to investigate the effects our solutions could have on commuters - how would the addition of these potential delays slow down traffic? These and other questions would be important to answer. Only then could we would expect our visualization to shift more towards making a final decision on the best solution. For now, we have set the basics out for a robust and important discussion about pedestrian safety and convenience in the Chester Square neighborhood.

Acknowledgements