Making Our Data Accessible

List of Sources

• Section 1: What's a Parklet?

  2019 Intercept Data

  “Citywide Assessment of Parklets & Plazas: Summary of data collected for Summer 2014 Public Life Study” (Panganiban and Abad, September 2014) (report) / (source)

• Section 2: Hyper-local Benefits

  2017 ACS 5 year (source)

  San Francisco Census Tracts and Census Blocks (source)

  San Francisco Office of the Assessor-Recorder’s secured property tax roll (2007 -2017) (source)

  San Francisco Affordable Housing Inventory (source)

  San Francisco Parklet Permits (source)

• Section 3: The Ripple Effect

  San Francisco Office of the Assessor-Recorder’s secured property tax roll (2007 -2017) (source)

  Assessor Blocks Geography (source)

  San Francisco Registered Businesses 2007 - 2017 (source)

• Section 5: It All Started With Feeding The Meter

  San Francisco Parklet Permits (source)

  San Francisco Parklet Database - Podio

• Section 7: What’s Next?

  Global Parklet Fusion Table (Robin Abad Ocubillo, San Francisco Planning Department with updates from Stae) (source) / (map)

  San Francisco Parklet Permits (source)

Data Analysis Methods

Section 1: What's a Parklet?

The data from "What are people saying" tab in this section are from intercept surveys carried out by Openbox and Stae in June 2019. These surveys took place at each of the six parklets whose sponsors were interviewed for this study and surveys were designed to approximate the city’s public life study standards manual, with the intent of helping us understand the sentiments of parklet users, passersby, and nearby business owners. The intercept instrument is available at this link.

The data from "What are people doing" tab is from a 2014 study designed by the San Francisco Planning Department to evaluate parklets and plazas. Data was collected in June and July throughout the week during peak afternoon and evening hours. The two survey instruments used in this 2014 study were pedestrian and cyclist counts, taken in 10 minute intervals twice an hour, and stationary activity scans, which categorize different types of postures and activities. The 2014 study helped to inform the of The Public Life Data Protocol (PLDP), which is a set of observation metrics that attempt to illustrate the diversity of activity in public spaces. PLDP is also a data structure for how cities collect and store public activity data. The standard was developed by Gehl Institute and its partners—Gehl, the practice, the Municipality of Copenhagen, the City of San Francisco, and with the support and input from Seattle DOT. PLDP is now open source and its survey instruments are publicly available on GitHub. All of the raw data are available to the public for the first time on Stae. For more information on this study and analysis, see the reports: Citywide Assessment of Parklets & Plazas (Panganiban, Ocubillo, 2014) and Opportunity Mapping (Stroman, 2014).

Section 2: Hyper-local Benefits

For this section’s Census data, our approach was to include census tracts that touch a neighborhood walkshed, which is an ⅛ mile buffer (interactive map of buffers) from the center of a parklet. An ⅛ mile buffer is the equivalent of a 3 minute walk. This is a similar neighborhood walkshed used in a public space study that included parklets called Opportunity Mapping (Stroman, 2014). We chose this buffer because we observed a parklet’s impact to be on a hyper-local, block-level scale. Many parklets are on the border of several census tracts and a buffer captures those neighboring tracts. There are a few outliers to this process, such as the parklets on Townsend in SoMA. We represented the demographics for these parklets within the census tract it sits in because of its location and proximity to the San Francisco Bay. Taking a ⅛ mile buffer here would overrepresent access to these parklets because much the census tract extends beyond a walkshed.

Section 3: The Ripple Effect

For the assessments, affordable housing, and business licenses, the ⅛ mile buffer is used as a geospatial filter. Any points contained in this filter are counted to approximate the “Ripple Effect” for each parklet on its block. Since these datasets contain actual latitude and longitude points, we are able to filter for discrete points within the buffer and then filter by year.

We used turf.js, which is an open source geospatial tool written in javascript, to create the buffers. See this live code sandbox on how this library works with data on Stae. After creating the buffers, we used them as geospatial filters for assessments and affordable housing units in Stae. We then took the sums of each annual assessment, utilizing R Studio to group individual assessments by each block. Finally, we took the sum and mean of each assessment block within the parklet walkshed and plotted them on a year over year line. The source code of this analysis in R is available at this GitHub repository.

Business Licenses were filtered within the parklet buffers using ArcGIS by Meital Hoffman, who is the public life intern at San Francisco City Planning.

Section 5: It All Started with Feeding The Meter

Data on parklet permits and installation dates and current status were pulled from the SF Planning Department Records. This data was reproduced manually and uploaded to Stae and is now available to the public in an open, machine readable format for the first time.

Section 7: What's Next?

Global Parklet data are from San Francisco’s Planning Department. This dataset was incomplete and therefore updated for this data story.

Qualitative Methods

Sponsor Interviews

We visited sponsors on-site at their parklets to talk about their personal experiences, motivations, challenges, and best practices.

• Time: 60-90 mins

• Sample size: 7 sponsors

• (See the discussion guide)

Parklet Documentation

We visited different parklets to observe their design, activity level, and neighborhood context.

• Time: 10 mins

• Sample size: 32 parklets

• (See the shot list)

Parklet Observations

We observed group sizes, activities, and interactions that took place in select parklets.

• Time: 20 mins

• Sample size: 6 parklets

• (See the observation guide)

Intercept Surveys

We talked with parklet users, passers-by, and local businesses to understand their impressions of parklets and the impacts parklets have on their lives or businesses.

• Time: 10 mins

• Sample size: 63 people

• (See the intercept guide)

Provotypes

We asked passersby and parklet users to write down their thoughts about what makes a public space feel inclusive or exclusive.

• Time: 2 mins

• Sample size: 22 people

• (See the provotype)

City Staff Interviews

We talked with members of the SF Parklets team about the program’s backstory, goals, challenges, and next steps.

• Time: 90 mins

• Sample size: 4 people