COHHIO HMIS

An open-source Homeless Management Information System (HMIS) custom reporting project

This repository would be helpful to any HUD-designated Continuum of Care looking for a way to get more out of their HMIS data. While there are some reports here, this repository's most important function is getting data from HMIS to a tidy form that can be used in other projects (and repositories). Think of it like a staging area for other projects.

About

This is an open source project released under the GNU AGPLv3 license. See LICENSE for more details or visit the official GNU page at http://www.gnu.org/licenses/agpl-3.0.html.

All the code in this repository is written using R in R Studio. Please consult the book R for Data Science for help getting started with R and R Studio.

Data Sources

1. The HUD CSV export - Every HMIS should be able to export this regardless of HMIS vendor.
2. Other data as necessary, to supplement what is available in the Export.

While your reporting needs may be different, this report contains data we need for our reporting that is not included in the HUD CSVs. These include but are not limited to:

• SPDAT Scores
• County (Enrollment level)
• Entry Exit Type (specific to ServicePoint)
• Users' Default Providers
• Full Provider and Organization Names (beyond the 50 character limit)
• Grant Type (specific to ServicePoint)
• Veteran data for Coordinated Entry (Custom fields)
• Housing Offers (Custom fields for Veteran Active List)

You may not need to supplement the data in your HUD CSV export, so whatever you use from this repository will need to be adjusted to leave it out. Or you may need a different set of miscellaneous data which will need to be incorporated to fit your needs. Other CoCs using this repository would need to work out what other data would be needed and how that data could be obtained.

If you want to know exactly what data I'm using, see below, and email me for more specifics if you need it.

Workflow

The workflow I'm currently using (which I expect to become less tedious eventually) is the following:

0. Be sure you have the following directories in your R project:
   • data (to keep your HUD CSV files and whatever other data you will use)
   • images (see #4)
1. Download and unzip your HUD CSV export into the data directory of the R Studio project. Permanently delete the .zip once it has been extracted so the PII doesn't remain on your hard drive.
2. If you're aiming to use the same exact ancillary data that I am, please see the notes below about what I'm pulling in and from where. You will probably want to build out your own data that's not available in the Export or you may find that the Export has everything you need for now. Either way, download any other necessary data into the data directory.
3. Run 00_a_script_to_rule_them_all.R, which will run all the other scripts that begin with a number. You can modify yours to fit your needs.
4. Doing 3. will create a .RData image file for each script that runs. Each image file has an extension of .RData and will be dropped into the images folder.
5. Your other projects (like R minor or R minor elevated) each have data folders with symbolic links in them, each pointing to the relevant .RData files.

This will one day be written as a package, hopefully.

I hope you can find useful code here. Please share feedback in the issues or email me at genelledenzin at cohhio dot org!

The Data I'm Using with the HUD CSV Export

1. Enrollment-level (at Entry): a. County Where Served b. County of Prior Residence c. Entry Exit Type (a ServicePoint-specific feature) d. This is all in an ART report named RMisc. I run it in View Mode and download it to .xlsx.
2. User data: a. Users' Default Provider (a ServicePoint-specific data element). We need this because we have a shared provider across the entire CoC and we need to be able to show where an Entry came from without listing the User's name. b. Same as what is in the Export, but currently the WS Export only includes "Active" Users, which doesn't work well for joining to the other data tables using UserCreated. We need this for Data Quality purposes. c. This is all in an ART report named RMisc. I run it in View Mode and download it to .xlsx.
3. Project data: a. Full Project Name (the Export only outputs 50 characters) b. Full Project Organization Name c. Grant Type (a ServicePoint-specific data element). We use this to section off certain types of grants, but honestly we could cut this and use the Funder data instead. d. Project County. Annoyingly, this is not a HUD-required data element so it is not in the Export and we rely on it HEAVILY, being that we are a Balance of State. e. This is all in an ART report named RMisc. I run it in View Mode and download it to .xlsx.
4. Case Manager data: a. A couple of our workflows require users to enter Case Manager data. b. I'm using ReportWriter for this. c. PersonalID, EEProvider, CMUserCreating, CMProvider, CMStartDate, CMEndDate
5. Custom Veteran data: a. I'm using ReportWriter for this. b. PersonalID, EnrollmentID, DateVeteranIdentified, ExpectedPHDate, HOMESID, ListStatus, MostRecentOfferDate, MostRecentOfferStatus, PHTrack, SSVFIneligible, VAEligible.
6. Project Evaluation scores for CoC Competition a. This year for the CoC Competition, the CoC team is entering Housing First scoring and such into each project's custom assessment we created and that data is coming out in the Project Evaluation reporting. b. This is coming from ReportWriter. c. ProjectID, ProjectName, ProjectType, CostPerExit, DateReceivedPPDocs, HousingFirstScore, ChronicPrioritizationScore, OnTrackSpendingScoring, UnspentFundsScoring. c. This is coming from ReportWriter.
7. Offers (part of custom Veteran data) a. PersonalID, AcceptDeclineDate, OfferAccepted?, OfferDate, PHTypeOffered b. This is coming from ReportWriter.
8. Referrals a. The Referrals available in the Export are limited to PATH-related ones only. We need all of them. b. PersonalID, HouseholdID, NeedID, ReferralDate, ProviderCreating, PATHReferralOutcome, PATHReferralType, ReasonReferralCanceledorDeclined, Referred-FromProvider, ReferralOutcome, ReferralsRanking, Referred-ToProvider, RHYMISReferralType, UserCreating c. This is coming from ReportWriter.
9. SPDAT Scores a. This is coming from three different ReportWriter, since the data is split into fam, ind, and tay. b. PersonalID, ScoreDate, and Score for each report.
10. Services a. Like Referrals, the Services in the Export are limited and we need all of them. b. This is coming from two ReportWriter reports, one for the basic info, and the other for the cost data. c. The main report contains: PersonalID, HouseholdID, NeedID, ServiceID, ServiceStartDate, ServiceEndDate, Code, Description, ServiceProvider, ProviderCreating. d. The costing one contains: ServiceID, Fund, Amount.

Security

No HMIS data is ever included in this repository. To make this code work, you will need to supply your own HMIS data. You are responsible for securing your HUD CSV export on your computer and ensuring that it is not compromised using whatever security measures you use for that locally.