Change agent data pre-processing

HSLdevcom / mal-effect-calculations

Impact assessment scripts for the Helsinki region MAL planning process

https://www.hsl.fi/hsl/mal

European Union Public License 1.2

1 stars 0 forks source link

Change agent data pre-processing #13

Closed attesn closed 3 years ago

attesn commented 3 years ago

Previously agent data has been combined in data pre-processing before plotting. However, this does not work if analysis is not based on predefined groups.

Move group/combine to plotting scripts and data pre-processing includes only cleaning and defining variables :

Change combine_agent_data. R to tidy_agent_data.R. Use Rdata files to save data frames for different scenarios.
Group agent data tables in plotting scripts

PR also removes car access from zones plotting as this result is no longer produced by Helmet-model.

attesn commented 3 years ago

Sure! Sorry I forgot that column names of input data had some changes.

I've later changed the Helmet input data to have two tables agents and tours. This requires further updates to R scripts. Would you prefer that I add those changes here as well, or should I make a new PR?

johpiip commented 3 years ago

@atte-wsp Hmm, you can add it here since it is further "changing agent data preprocessing". :) I'll review this once you've pushed new commits and uploaded the V2030_ve0_rm data.

attesn commented 3 years ago

@johpiip Helmet data is updated to Sharepoint folders and changes are added here. I've been using V2030_mal19_v1 environment in config file.

attesn commented 3 years ago

@johpiip Seems that some refactoring of code is still needed as calculation is quite complicated and slow after latest changes. I'll comment here when done.

johpiip commented 3 years ago

No problem, I will wait until you are done, @atte-wsp! :)

johpiip commented 3 years ago

Hey @atte-wsp, sorry to bug you, but I glanced through some of your code. As a pre-review comment, could you Ctrl+Shift+F for library(sf) calls, and edit them so that you first call library(tidyverse) and then library(sf)? And could check that the results stay the same? sf docs recommend that particular order: https://r-spatial.github.io/sf/reference/tidyverse.html If the coordinates cause problems, you can sf::st_drop_coordinates() before using dplyr verbs.

attesn commented 3 years ago

Data pre-processing is ready for review. I'll add another branch for plotting as there are tons of updates coming with new effects and removing / updating old ones.