Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Logan
    @logantpowell
    the highest-resolution (smallest geographic areas) are going to be the sf1 (decennial), but that's older data (2010)
    a balance between would be the acs 5
    Elicere Inc
    @ElicereInc_twitter
    can anyone help point to a nice, clean variable list of all variables (i believe there are over 4700) that can be queried in CitySDK ?
    Logan
    @logantpowell
    Hi there, you can find the full list of available variables here: https://api.census.gov/data.html
    if there's anything in there that you don't see in CitySDK, we can add it for you. Just let us know!
    Logan
    @logantpowell
    @atsepkov sorry I missed your last message. Let me know if you have some time later this week to do a screenshare
    Alexander Tsepkov
    @atsepkov
    @logantpowell np, I worked around that particular issue already. On a related note, CitySdk accepts a zipcode as input. However, the output in that case seems to use the data of closest "city" instead of a zipcode, which is often wildly different. For example, if I try 90210, I get population of 34,663 rather than 22k. For inner-city zip code it can get even worse. For example, 60620 (a neighborhood in Chicago with population of 72k) is returning me what seems like the population of ALL of Chicago (2,717,534). After doing some research I realize that census doesn't measure in terms of zip codes but is there any way to approximate the regions using tracts to obtain more accurate population counts and more relevant demographics? The tract-to-zip mappings are available for free (https://www.huduser.gov/portal/datasets/usps_crosswalk.html) but given the size of these files and a requirement for a separate census query for each tract, this is not something that would scale on my local machine. if census server allowed bulk request of multiple tracts at once (or had a concept of a zipcode) that would help a lot). Is this something feasible for the project?
    Logan
    @logantpowell
    the way the citysdk uses zipcodes is a bit different than the hud files... we take the centroid of the zipcode just as a way to focus your geographic request: for example, if you want all the tracts in a specific county you can use any zipcode in that county as the location; county as the 'container' and tracts as the 'sublevel'
    we're considering incorporating datasets that leverage crosswalk data such as the one you've referenced from hud (thank you for that, btw). There's also this one: https://s4.ad.brown.edu/projects/diversity/Researcher/Bridging.htm if you're interested, but for the time being we are without a developer and will need to hold off on any new features... if you would like us to include this, you could build it and submit a pull request with a schema and database for us to host?
    Alexander Tsepkov
    @atsepkov
    @logantpowell can I control the sublevel, I thought it only takes true/false as input, or is tract automatic sublevel of county?
    Alexander Tsepkov
    @atsepkov
    also, for some reason the huduser website I linked seems to be overly generous with assigning tracts to zipcodes, for 90210 for example, I'm getting over 45k people. The mappings on http://proximityone.com/ziptractequiv.htm seem a lot more accurate (and exclude 5 tracts huduser seems to include. In the end I was able to get 25k for 90210 and 72k for 60620 using proximityone tract mappings, which is much more accurate, accurate enough for average user in fact. Problem is, their map isn't free, but $495 for the downloadable version
    Alexander Tsepkov
    @atsepkov
    actually, it seems like sublevel per-tract populations returned from citysdk don't match up with other sources
    for example this tract data from census:
    { name: 'Census Tract 1439.02, Los Angeles County, California',
    state: '06',
    county: '037',
    tract: '143902',
    population: '3003' }
    whereas http://proximityone.com/ziptractequiv.htm claims that tract's population is 82, and proximityone data actually adds up to correct zipcode data, so is citysdk wrong or am I doing something wrong?
    Logan
    @logantpowell
    @atsepkov you're right: sublevel should be 'true' and level should be 'tract'
    our zipcode implementation is rather rudementary, being hard-coded (here)[https://github.com/uscensusbureau/citysdk/blob/master/src/resources/zipcode-to-coordinates.json]
    I will start a dm with you...
    Tom Arnold
    @tomarnold2
    Hi, everyone. I'd like to be able to confirm if an address is within a city's or county's borders. I searched through TIGER on census.gov and also on Google Maps to see if there was an API for this (or a way to write my own) but I've had no luck. I've spent a few hours researching before bugging you good folks, but now I'm at the point of asking for help since this cannot be a rare request. Any advice where I should look to find this existing functionality or to at least find the building blocks to create it? - thank you.
    Try that
    Tom Arnold
    @tomarnold2
    Thank you, @logantpowell - if I'm reading that correctly, though, that's for looking up individual or batch addresses. I want this, for sure, but I also need a way to determine if that address resides with a city's limits or even a particular county's boundaries. (For example, my address is Wenatchee, Washington, but I'm outside of its city limits).
    Logan
    @logantpowell
    Just trying to understand what you're looking for is a way to verify whether or not an address is in a particular area?
    Logan Powell
    @loganpowell
    when I say particular, I mean predetermined
    BenTheMonk
    @BenTheMonk
    Hi everyone, I am new here
    Logan
    @logantpowell
    Hi Ben!
    BenTheMonk
    @BenTheMonk
    Howdy, Logan
    Tom Arnold
    @tomarnold2

    Hi, @logantpowell - here's my example: a government customer sells a type of city license on their website but it's only necessary for people who live within city limits. Too many people don't know this and attempt to buy the license even though they live outside of city limits (but, they have the same city name and postal code). So, I cannot use postal code alone to determine if they are in city limits. The website needs to alert the user that they are not required to make that purchase since they are outside of city limits (otherwise, the city has to refund those purchases and it's a hassle).

    The point is that if you look at Google maps for a city, you can see the boundary lines for the city limits. So, the data is out there. What I'm trying to figure out is how to check if an address is located within city limits and zip code alone won't do it. I need to get the lat/lng for the user's address, then check to see if it's in the city limits, then warn them that they don't need to do the purchase if they are outside of city limits.

    Sorry I'm not being more succinct. :-) I hope that helps, though?

    Logan
    @logantpowell
    If what you trying to do is "given a geographic area, does this user fall within its boundaries type of thing you may have to do some postGIS type stuff
    Check out terraformer.I'm
    Terraformer.io
    Cameron Banowsky
    @cbanowsky
    Hey anyone have a sample app or something similar to isolating education and age based on region or zip?
    Logan
    @logantpowell
    Cameron, we have education and age in ACS, which - in the 5-year collections - gets you down to zip
    Cameron Banowsky
    @cbanowsky
    omg could you share?
    Logan
    @logantpowell
    are you just looking for stats or would you be creating a map?
    Cameron Banowsky
    @cbanowsky
    save me so much time. I'd like to create map but data first :)
    god I love this community already
    Logan
    @logantpowell
    so, the citysdk might help which combines stats with geographic shapes/geojson, but if you want just the stats I can share the underlying stats api call for that
    which would you like?
    sorry wrong link sec
    Cameron Banowsky
    @cbanowsky
    Yeah that would be great, I will chart myself. Thank you so much
    Logan Powell
    @loganpowell
    this gets you total population (B01001_001E) males (B01001_002E) and females (B01001_002E) for all zipcodes in the US
    you can find a complete list of variables for ACS 5 (2015 latest vintage) here: http://api.census.gov/data/2015/acs5/variables.html
    Cameron Banowsky
    @cbanowsky
    thank you @loganpowell seriously saved me a couple hours
    Logan Powell
    @loganpowell
    you're very welcome
    Cameron Banowsky
    @cbanowsky
    i will be lurking in this channel lol
    Logan
    @logantpowell
    haha
    Cameron Banowsky
    @cbanowsky
    @loganpowell how can I use tableau? This is a new program to me. Import json or other?
    Logan
    @logantpowell
    sorry, i didn't mean to add the tableau reference
    but you can use it