Iss435 #437
Iss435 #437
Conversation
still working on CT and quality checks
|
The county homeless student files for 2019-2022 are ready. Specifically the total homeless student file and the subgroup by race file. One big flag: CT counties prior to 2022 are not in crosswalk and therefore not in this data. This will need to be rerun when the crosswalk is fixed. |
Add folder for final forms
Added 2014 PEP population data into the crosswalk manually since the API is limited
…gh 2021 As discussed, aligning with the data team decision to maintain the original 8 CT counties through 2021. Also manually adding PEP population data for 2014 to complete previously missing data.
Kicked off more specific documentation about the crosswalks in the README - will revisit as needed
|
CT is now available for prior years based on the updated crosswalk for all homelessness 2019-2022 as well as subgroups 2019-2022 |
|
@ekgutierrez1 just finished reviewing. I will have line item comments in a bit. I need to do another pull to update for the commit you just pushed. |
There was a problem hiding this comment.
@ekgutierrez1 added line item comments! Many of the comments from the city code also apply to the county as much of the code is duplicated.
| cap n ssc install libjson | ||
| net install educationdata, replace from("https://urbaninstitute.github.io/education-data-package-stata/") | ||
|
|
||
| *Set up globals and directories |
There was a problem hiding this comment.
In the checklist on the repo Wiki under reproducibility there is " The program runs from start to finish without stopping due to errors or incompleteness". This currently does not pass because of some of the code here in the set up section. The code will error if the folders in the mkdir lines already exist. I'm not sure how to do this off the top of my head, but the code should check to see if they exist first and then mkdir if they don't.
There was a problem hiding this comment.
The error it provides (that the folders already exist) does not prevent the rest of the code from running. If it is a new user, this code auto-creates the paths they need. Since it doesn't prevent the rest of the code from running but is necessary for the paths of the code, I'd prefer to leave as is (this is how all of my files currently run, but we can ask the higher ups).
| clear all | ||
|
|
||
| global gitfolder "C:\Users\ekgut\OneDrive\Desktop\urban\Github2\mobility-from-poverty" | ||
| global years 2019 2020 2021 // refers to 2019-20 school year through most recent data | ||
| global gitfolder "C:\Users\ekgut\OneDrive\Desktop\urban\Github\mobility-from-poverty" |
There was a problem hiding this comment.
Also from the checklist: "The program avoids hardcoding local file paths and instead uses global paths that will work regardless of where the program is being ran"
There was a problem hiding this comment.
This uses a global - the user will have to hardcode this upon download. There isn't a way around this unfortunately. That's just how globals work.
| ***************************** | ||
| ****City/Place Crosswalk***** | ||
| ***************************** | ||
| ** Import city crosswalk file to edit names of city crosswalk to match city location strings in CCD school district data |
There was a problem hiding this comment.
It would be helpful if comments described more specifically what is happening. For example it looks like you are converting the place codes from numeric to string, then adding leading zeros but that isn't really described in the comment.
There was a problem hiding this comment.
I added relevant comments to this and the county dofiles.
| } | ||
| save "intermediate/ccd_lea_recent_city_race.dta", replace | ||
|
|
||
| *merge two ccd datasets together |
There was a problem hiding this comment.
It looks like the merge has 3,896 unmatched records. Is that expected? It would be helpful in the comment to describe what the expectation is.
There was a problem hiding this comment.
Yes, it's expected. I added a line and a comment to explain.
| unzipfile "EdDataEx Homelessness `year'.zip", replace | ||
| } | ||
| cd "${gitfolder}\02_housing\data" | ||
|
|
||
| *import csvs | ||
| *Due to changes in EdDataExpress website, 2022-23 data must be manually downloaded. Please follow the following steps. |
There was a problem hiding this comment.
This causes me a lot of stress. When I went to the website there is a big red banner that says
Due to current system issues, datasets must be limited to fewer than 150,000 rows. A selection of data greater than 150,000 rows may result in a truncated dataset.
It looks like we may need an export larger than 150k so this may be a problem. Aside from that, it requires an amount of human intervention that is prone to error.
I don't know what the best solution is if the file is not able to be downloaded programmatically. Perhaps @awunderground or @jwalsh28 can weigh in?
There was a problem hiding this comment.
I worked this out with Claudia as the best case solution for this - but welcome other comments. Their website is problematic right now, so we gave instructions as best as possible in the comments for how to download the data. If you follow the instructions, it should work without the 150,000 problem.
| } | ||
|
|
||
| *new as of 4/13/23 - updated 2/8/24 - in ACS-based metrics, if it was less than 30, it's set to NA | ||
| *if aggregated enrollment is less than 30, quality of the variable is 3 | ||
| replace homeless_quality = 3 if enrollment<30 |
There was a problem hiding this comment.
Should the replace homeless_quality = 3 if enrollment<30 line be part of the loop?
There was a problem hiding this comment.
No, the next set of code/loop takes care of the individual race categories. It total homeless student count is less than 30, regardless of the individual race counts, the quality should be 3. However, if the individual race categories homeless counts are less than 10, they should be replaced with NA, because they are smaller categories than aggregated counts.
| replace `var'_count=-1 if `var'_count<=2 | ||
| } | ||
|
|
||
| *merge to crosswalk of places/cities | ||
| merge 1:1 year state city_name using "intermediate/cityfile.dta" |
There was a problem hiding this comment.
We are losing a vast majority of cases on this merge? Is that right? Felt like a red flag to me.
There was a problem hiding this comment.
It is correct because we only care about matches 2 and 3. I added some comments in the dofiles to help clarify.
|
|
||
| *check quality | ||
| **************** | ||
| *Quality Checks |
There was a problem hiding this comment.
A lot of quality checks are difficult for me to interpret as someone without much topical knowledge of the data. For checks where a large number of summary stats are created it would be helpful to know what to look for in those numbers. Alternatively using pass/fail checks would be a bit easier to interpret.
There was a problem hiding this comment.
Added assert commands where relevant and comments for other checks.
| *we replace other==. to other==1 to mirror lines 147 in the other 4 race categories | ||
| replace other = 1 if other==. & year==2019 | ||
| *we replace other==. to other==1 to mirror other 4 race categories | ||
| replace other = 1 if other==. |
There was a problem hiding this comment.
There are two lines of commented out code in the loop:
*replace var'="1" if var'=="S"
*destring `var', replace
Is this temporary or permanent? If permanent please delete these lines. It doesn't look like this is commented out in the city code.
There was a problem hiding this comment.
Great catch, I have now deleted that commented out code.
| tab `var'_quality if `var'_share==. | ||
| tab `var'_quality if `var'_count==. | ||
| } | ||
| *To reviewer: in 2019, state 17 and county 061 there is an instance where black enrollment is 0 but homeless count is 3. |
There was a problem hiding this comment.
If we think this is an issue caused by the underlying data then it should probably be switched to the most conservative (unreliable data) flag and documented somewhere
There was a problem hiding this comment.
It's already a flag = 3 (most unreliable),but based on the underlying data, I think it should actually be NA. Making this change now.
I separated the three subgroups from ELA subgroup data it three separate eval forms. The "ela_subgroup_county" file should be deleted "
Fix final evaluation function
The city homeless student files for 2019-2022 are ready. Specifically the total homeless student file and the subgroup by race file. I'll now turn to the county versions of the same.