Investigate SMMI typologies

[annacd114]

Look at all the SMMI tracts for the Bay and see which attributes of more exclusive tracts they do have vs. not, essentially why they are stuck in SMMI if they almost are exclusive

[annacd114]

(one sheet of all SMMI, Tract ID then columns for all relevant displacement typologies: AdvG, ARE, BE, SAE. For each tract, identify how many boxes the tract ticks for that other typology & indicate out of how many boxes there are to tick)

(another sheet: ARE with all its parameters, then a 0 or 1 if that tract matches that parameter. then sum the rows)

(same for the other 3 relevant typologies)

do this first for SF, then Seattle - see how similar the proportions are for the two regions, summarize findings and discuss with Tim. Use the "typology output.csv" --> soon!

[annacd114]

My findings:

SF --> https://docs.google.com/spreadsheets/d/1Zu_o3KF3TlQ-1gjLCzqTBHwu_0WKX5_3/edit#gid=583126133

Seattle --> https://docs.google.com/spreadsheets/d/1CkzHUkjL3bLxnHpxch3EqRVmSA8av7Du/edit#gid=468605918

[annacd114]

Preliminary notes:

• Nearly all SMMI tracts had a 'rapid increase' in housing costs. This excluded nearly all of these tracts from Advanced Gentrification and At Risk of Exclusion, even though many of them met the income level and affordability metrics for these types. However, they could not be Becoming Exclusive largely because few tracts checked the "loss of low-income households" box, and this category was tough to get into with six criteria to meet. And they could not be Stable/Advanced Exclusive because 0 of these tracts in the "affordable to predominantly high-income" category, for both sites, even though all of these tracts fell within the change in housing cost options for this type.

• Very few tracts met the criteria for having 'gentrified,' which barred them from Advanced Gentrification. I noticed that the "dense" column was an interesting component of this pattern. For example, some tracts did gentrify in some fashion but didn't count in the typology because of whether or not they were dense. And in Seattle, 0 SMMI tracts were dense!

[annacd114]

Anna to send proposed approaches based on this analysis out to the group by Friday

[timathomas]

Copy of @annacd114 email: Advanced Gentrification Moderate, mixed-moderate, high, or high-income Affordable to moderate, mixed-moderate, high, or mixed-high income Marginal change, increase, or rapid increase Gentrified

Becoming Exclusive Moderate, mixed-moderate, high, or high-income Affordable to moderate, mixed-moderate, high, or mixed high-income Rapid increase Median income change Loss of LI OR declining in-migration (not both)

Stable/Advanced Exclusive High-income, 2000 and 2018 Affordable to high or mixed-high income Marginal change, increase, or rapid increase

—— What if we said that AG does not have median income change and SAE does not have loss of LI or Declining in-migration? I think that would make them mutually exclusive. Thoughts @profkdc and @annacd114?

[annacd114]

I think you might need to exclude both of those things for both types for it to be mutually exclusive. The duplicates all have both median income change and loss of LI/declining in-migration.

[timathomas]

This is hard! I’m still struggling with how to fix these mutual exclusivity. Changing the values as I mentioned can’t really be done, it goes against what these categories represent - e.g. the opposite of loss of low-income households is a gain and doesn’t work with AG. I’m starting to think that maybe if there are two pairs, we choose the highest of the two: AG to BE & BE to SAE. Still playing with this.

[profkdc]

This is exactly how we handled it before (we have had overlaps in other categories historically, particularly AG and exclusive). We prioritize AG and go from there. I think it's okay though I understand your conceptual discomfort!

On Mon, Aug 31, 2020 at 1:17 PM Tim Thomas notifications@github.com wrote:

This is hard! I’m still struggling with how to fix these mutual exclusivity. Changing the values as I mentioned can’t really be done, it goes against what these categories represent - e.g. the opposite of loss of low-income households is a gain and doesn’t work with AG. I’m starting to think that maybe if there are two pairs, we choose the highest of the two: AG > BE & BE > SAE. Still playing with this.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/cci-ucb/displacement-typologies/issues/39#issuecomment-684014318, or unsubscribe https://github.com/notifications/unsubscribe-auth/AOKT6AVDDF75E2B4SONP32TSDQAL7ANCNFSM4QGQ254A .

[timathomas]

Need to deal with the AG, BE overlap.

[timathomas]

Now issue #64