Data from https://data.cityofnewyork.us/City-Government/DSNY-Grafitti-Information/gpwd-npar

NYC OpenData

New York is one of the United States leading proponent for open data and government transparency. The city has stood by the Open Data movement from the beginning, initially gaining public acclaim during President Obama’s first term. The movement surrounds the practice of publicizing data collected by citizen funded departments. Data varies in content and form, from financial balance sheets to statistical records. In the past year, emphasis has been made to release information into machine-readable formats: sans-PDF, pro-JSON.

CivicHacking

Yesterday I received an email promoting the Code for America brigade supporting the “Code Across NYC” civic hackathon. The email, from Chris Whong, promoted the spirit to getting involved with local communities through any means. Following Chris, Noel Hidalgo shared a recently released NYC OpenData blog post on new real time datasets on New York city.

The data set topics ranged from medical history, farmer’s market records, and graffiti reports. All the sets were deemed “real time” and machine-readable. These data sets provide public access to rich repositories of information. Based on the availability, these will be the beginning of fresh new software applications based on consuming these datasets. Imagine the next farmer’s market notification mobile app or graffiti report location tool.

Individually, the data makes available the well-organized datasets. Even more exciting are the cross-reference opportunities. An interested party doesn’t need to know how to program. Simply grab the CSV export copy of a preferred data set and get to some analysis. Immediately a PivotTable reference between data and category can provide interesting high-level comparisons. Looking at graffiti creation by date across boroughs shows the uneven reporting of graffiti incidents.

Graffiti

To start, I love graffiti. I love the notion of people finding and accessing places that have been deemed ‘dead’. I love the raw-ness of spray-paint. I love the culture, the people, even the drudgery associated to the painters. Most of all, I love how it represents a community of people who desire to express themselves publicly.

I am not one to argue that graffiti is not vandalism.mi strongly disagree with the painting of private property. I instead believe graffiti is the effect of a greater social culture associated with stemming the voice of individuals. I believe that large amounts of graffiti in an area are public signs for improvements.

There is a person doing the painting and that person is trying to say something. Maybe they have poor handwriting and a poor value system, but they are people to understand. Why do they get pleasure out of scrawling their moniker in public places? Why do they write on things that don’t belong to them? These questions make me interested in drilling into the available graffiti dataset.

I have a hypothesis that areas that are not central urban sectors with high amounts of graffiti could be improved by providing public art programs. I think most of the periphery communities with large amounts of graffiti are the result of youth between the ages of 12-18. I also believe these students have poor grades in school, are high risk for dropping out, and are likely exposed to drug use.

These assumptions aside, the graffiti is an expression of a societal problem. It’s not a graffiti problem. It’s. Community issue. It’s an education issue. Graffiti is the signal, but the cause is larger.

Action

I have never been able to pinpoint or analyze graffiti from a statistical front. I have never seen a dataset like the one that NYC has made publicly available. I’m sure these sets have been stringently analyzed and reviewed by police and sanitation departments, I believe the information was most likely viewed from a cost-arrest standpoint. I believe the individuals who, like myself, love graffiti and see the value in its messaging will make great use of this information.

I will be looking at this information to seek out insight for potential public programs to stem these issues. Imagine, graffiti declines because citizens feel alternative ways to express themselves. Imagine, communities coming together to take responsibility for the visual composition of their neighborhood. Imagine, children and teenagers out on the street with a spray can, because they feel they are being heard.

Algebra for designing layouts

I have been reading about machine learning and linear algebraic analysis. In college, I failed two calculus courses. In high school, I repeatedly failed higher-level algebra courses. In an academic setting, pumping away at math worksheets was never a convincing use of my time. I much rather play video games at the time. I didn’t realize at the time, but I was using complex equations to calculate the optimal strategy to win.

Although I never caught on to academic mathematics, I find myself regularly using algebraic equations at work. I use complex algebraic equations everyday while design. While visual design is driven by aesthetic choices, the execution of interactive layouts requires thought out calculations. To get a navigation item to “stick” on a webpage when scrolling, or “unstick” after scrolling a certain distance, calculations are necessary.

After recognizing how something pleasantly interactive requires mathematically complex planning, I decided to explore other pursuits. I love math when its applied to something practical. I thrive in solving animation related math problems. Through reading about machine learning, I found a new topic that seemed interesting.

I attended the Soka Education Student Research Project conference this weekend. In preparation, I compiled the past 10 years of published articles. The compilation was made for WhoWroteForSESRP.¹ The process resulted in a simple JavaScript table for users to search published authors, titles, and issues.

Presenting data

The context in which you view Data (with a capital D) determines what you can understand. Edward Tufte emphasizes this point. I tried to understand the past 10 years of SESRP published papers. First, I organized the information in a flat context with simple attributes. Second, I viewed the relations between the attributes in aggregate. Lastly, I created charts and tables to make the results clear. This process shows that there are still many interesting relationships among SESRP papers that have not been clearly identified.

The first step required me to review the past 10 years of published journals. I selected the individual published paper titles, associated authors, and published issue year. The attributes chosen could have been expanded to full article texts, human-defined tags or categories, and article bibliography. Initially, I wanted to create a basic repository for easily browsing the past-published articles. ²

Second, I isolated the individual attributes to view them in relative increasing or decreasing order. Through collecting the data, I saw a distinct increase in published papers during the year 2011. Similarly, viewing the authors in aggregate clearly displayed the writers who published the large number of articles over the past 10 years. ³

Lastly, I charted the data to further understand the relationships between the authors, years published, and quantity of articles. In the future, I will add attributes to compare against. Mainly, I am interested in charting network relationships between cited articles, cited authors, and unrecognized topic driven clusters. ⁴

Insights from analysis

Filtering the quantity of articles published by year showed a unusual spike of submissions and a potentially changing overall trend. In 2011, the number of articles submitted more than doubled from previous year. In 2010 nine (9) articles were submitted, followed by 22 articles submitted in 2011. In the years following 2011, the average number of articles submitted increased by over 20%.

The quantity of articles published by author shows a familiar trend in participatory communities. Out of 85 total authors published in the yearly booklets, 64 authors had one paper published, 13 authors had two papers published, 3 authors had three papers published, 3 authors had four papers published, and 2 authors had five papers published. These numbers show that in the past 10 years from 2005 to 2014, approximately 75% of writers only publish one article.

The two authors who published the most are Gonzalo Obelleiro and Ryan Hayashi. Obelleiro’s first article appeared in 2006 and Hayashi’s first article appeared in 2011. While Hayashi and Obelleiro have the same number of articles published, Obelleiro has presented articles that have not been published, therefore technically he has written more than any other author. Regardless, Hayashi has been writing for less than half the number of years, so his rate of appearance is impressive.

Suggestions

The past 10 years of conference organization has been completely led by students. The proposal request process for SESRP papers is distributed through current student body, the alumni network, and individual requests. In the interest of revitalizing previously interested attendees, past SESRP paper submitters should be personally requested to submit new papers. Future student tasked with organizing the SESRP conference should seek to mobilize the 75% of one-time publishing authors to propose a new paper topic. Most past authors are alumni of Soka University of America, so the alumni network may be useful in gathering contact information.

As the 10-year anniversary booklet is being prepared, organizers should seriously consider the 25% of authors who published more than one paper. The eight authors who published more than 3 articles each should be consulted for advice. While quantity of published papers does not reflect a priori for understanding Soka Education, their insight should be considered. These authors have shown personal interest through their multi-year commitment. These authors, from most published to least, are:

• Gonzalo Obelleiro (5)
• Ryan Hayashi (5)
• Maria Sanchez (4)
• Jean Marcus Silva (4)
• Michael Strand (4)
• Masahiro Kaleo Louis (3)
• Menelik Tafari (3)
• Nozomi Inukai (3)

Future expectations

The three attributes listing of the SESRP papers from 2005 to 2014 have raised positive results. This method of organizing information and finding relationships will be further applied to understand the relationships between sources cited in submitted papers. The goal will be to find the unseen relationships between the submitted papers. The cited resource’s author and publications will help create rich network maps displaying relationships between SESRP authors. Similarly, this relational mapping may display under recognized clusters or categories. This will prove to be useful in the 10-year anniversary booklet’s drafting.