Interactive East Liberty rent data

Tech4Society is working on creating interactive and engaging data visualizations of census and ACS data. We want you to find and understand the dynamics affecting your city and your neighborhoods so you can successfully lobby for important policy changes.

Data

Each color represents the number of people paying rent within an interval of $50, $100 or sometimes more at higher rent values, e.g. above $1000. Cooler colors, at the bottom, represent lower rents, and higher colors, at the top, represent more expensive rents.

Interaction

When you click on an interval, the visualization will display how many people were paying up to that amount, for both 2011 and 2016, while connecting the two to highlight the trend. To reset the visualization, click on white space near the graphics.

Where the data comes from

This is American Community Survey 5-Year Estimate data for 2011 and 2016. We use the columns that indicate total number of renters paying gross rent within intervals of $50 and $100, e.g. between $500 and $550, or between $800 and $900. These buckets are unfortunately not available as adjusted for inflation, so 2011 dollar values are slightly undervalued compared to their 2016 counterparts.

This data is static, and obtained for Pittsburgh’s East Liberty, as defined by census tracts 1113 and 1115. It also only keeps track of renting units, not owned units.

Future work

At the moment, all data is static, and only for Pittsburgh’s East Liberty (tracts 1113+1115). In the future, you will be able to produce this visualization for any group of census tracts in the USA. This will allow you to look at rent trends for individual tracts, neighborhoods, cities, counties, or any combination you can imagine.

ACS Column specifications

The columns used for 2016 were B25063_003E through B25063_026E, and for 2011 B25063_003E through B25063_023E. The 2016 data has more buckets than the 2011, so we collapsed the last buckets of 2016 into a single one. The specification of the columns match otherwise.

2 Responses

  1. Christine Mondor

    Questions and comments:
    How does ACS calculate their number of units… is it based actual number of responses or are the responses projected across some other public database that shows an actual count of units? This graphic looks like there was no net change in total units, which is unlikely. Should the graphic represent percentages instead of number of units?

    Adjusting for inflation would be very helpful.

    If you were going to use this data to make the case of affordablity, how do we deal with the common condition of roommates? These tracts contain many mutibedroom units that host students and other unrelateds. It would be informative to understand self reported rents on a per person basis in nonrelated households. Or at least understand the percentage of units within each strata that are occupied by nonrelated roomates. Does the ACS survey includes a question to allow this disaggregation?

    • João Martins

      Thank you for the comments! There are indeed some limitations to using ACS data, as well as some great features. In the future, we will add a section about how reliable the data is, what to be careful with, etc.

      ACS 5-year Estimate data comes from a rolling questionnaire, and despite being the most accurate of the 1-, 2- and 5-year estimates, there are very significant statistical error margins for smaller areas such as neighborhoods. As such, these graphs are more indicative of general trend than exact demographics.

      Where ACS shines is in the ability to obtain this data dynamically for any place in the country, at any scale, with smaller error margins the larger the area.

      These following links explain the nature of ACS surveys, and compare the different 1, 2 and 5 year estimates:
      https://data-planet.libguides.com/ACS (left sidebar)
      https://www.census.gov/programs-surveys/acs/guidance/estimates.html (comparison between different estimates)

      For neighborhoods changing as fast as East Liberty is, we should expect 5-year estimates to lag far behind what is actually happening, e.g. many responses for the ’16 results will be from ’12, ’13, etc.

      Inflation adjustment changes the “buckets” themselves, e.g. $322.74-$376.53 for the ACS $300-$250 bucket ’11 in ’16 dollars (according to https://www.bls.gov/data/inflation_calculator.htm). We agree it is worth thinking harder about how to incorporate this into the graph, since it is important, although definitely not trivial.

      A quick perusal through the columns revealed some data that may be useful regarding roommate issue you raised, like columns B25031 and B25066 (see https://www.socialexplorer.com/data/ACS2016_5yr/metadata/?ds=ACS16_5yr). It would be great to add this information to a more complete infographic.

      We would be happy to collaborate with you to determine what data is important to advocate for good policies that counter the affordable housing crisis in Pittsburgh, and integrate that information into these visualizations. Furthermore, if the City collects more accurate data for Pittsburgh, as stipulated in the Resolution of Concern, we could tailor some visualizations to that as well. If you’d like to talk or meet, please contact us at cmutech4society@gmail.com!

Leave a Reply

Your email address will not be published. Required fields are marked *