Horizon Round 2 Model

This version contains all the additional code, data, and scenario settings used to implement the second round of the Horizon Futures. The largest change is that various policies were coded and set for application.

Horizon Round 1 Model

This is the version of the model that was used to construction the three Horizon Futures without any policy modifications. While it retains the TM2 outputs it also provides the outputs needed for TM1.5 (ie new post-processing but 1454 TAZ).

TM2 Outputs and Hazards

This version includes codes to produce model output and demographic post-processing for the new TM2 zones. It also includes sea level rise and earthquakes (though the latter is turned off here). Tenure has also been disabled.

Cleaner Version of the PBA40 Model

This release incorporates cleaning, refactoring, and commenting performed in the UDST and UAL directories. Data cleaning for the base year is moved into a preprocessing step. A number of other minor but substantial adjustments result in code that is no longer identical to the PBA40 version.

Major cleanup

These are major changes which make up some code debt from the last release. This should clean up and document a great deal of functionality from the last Plan.

See #70

Here is a relatively complete list (but I might have missed one or two points)

• left files on s3, but got rid of Makefile. Fetching files is now a mode of baus.py

• Got rid of vagrant file

• Got rid of data_regeneration scripts - will start soon on redoing these, but since they're preserved in the history no need to keep them in master

• Got rid of qgis maps

• Moved additional_units model to auxiliary_units model (renamed)

• Commented all code and settings.yaml

• Rewrote README for all these changes

• Moved run.py to baus.py to have a more specific name

• Reviewed scripts directory and added some description of what the scripts do (in the README) reviewed the scripts and got rid of those scripts we won't use again

• Now have preprocessing mode of baus.py so deed restricted units don't get assigned on the fly (saves time)

• Don't have baseyear simulation instead we write out base year summaries every time. We don't need to run the whole model system in the base year, and baus.py is modified to take care of this

• Proportional jobs model assigns only to the correct types - there is still some work to do here

• Homesales dataframe not required since redfin prices are assigned to buildings dataset

• TAZ variable computations were somewhat sloppy before - they've been rewritten

• use getopt in baus.py so now we have a pretty usage statement

• double check all caching settings

• set up travis to do pep8 and unit test checks - tests are modest at this point - I have a list of testing improvements to make

• we don't use numbers for building types - they're not two letter characters - HS, HT, HM, etc

EIR Alternatives

These are Mike's changes for the latest run of the scenarios which will serve as the EIR alternatives. This also includes scenario 5.

r7224 Final Draft of Preferred

Hundreds of commits to respond to feedback from stakeholders on the draft preferred

r5080 release new DRAFT of Preferred

Includes 100s of changes which are mostly used to specify the DRAFT Preferred

Nearing release of DRAFT "preferred"

Added a number of features including capping growth and adjusting vacancy rates in the base year, disallowing retail in certain counties, adding static parcels and zones to keep certain sectors of jobs in place (e.g. gov't jobs at city hall) and a proportional jobs model for when buildings get added to static parcels.

A number of specific changes that @mkreilly made in the scenarios.

Stable version used in workshops

I can't make an exhaustive list of changes since the last release since there have been 800 commits since then. I promise to do a better job of keeping release notes going forward.

In sum, this model runs the 4 scenarios used in the workshops, which will be documented in the baus_docs repository. The major innovations here are around the account system which is used to subsidize housing in sustainable locations and/or to fund affordable housing.

The hlcms are segmented by income quartile and the elcms by sector. The hedonics include a number of accessibility variables and neighborhood demographic variables.

There are a huge number of outputs, including summaries at the TAZ, jurisdiction, and superdistrict level, which are viewable in map form as well. There are also parcel level debugging csvs for all new developments that come out of the simulation, and there is a much larger csv which includes all buildings in the base and horizon years which provide the most disaggregate information for debugging.

Hooray for release 0.2!