For several years, we have collected Harris County eviction data from the courts every day, made it available on a real-time dashboard, and published several analyses of the data.
Harris County has some of the best eviction court data available in the country, and as a result, we have been cleaning and sharing the data with a diverse group of researchers.
This eviction data dictionary is a companion to the actual datasets, and it is designed to explain the data field by field.
There are five datasets that we use to complete our analyses:
- Eviction case metadata
- Enriched eviction case metadata
- Eviction docket scheduling data
- Eviction case events
- Eviction case documents
This post is designed to help those researchers understand what data is available, what kind of questions the data can answer, and where it falls short.
This is an unofficial data dictionary, and it was compiled without oversight from the court. If you use this data dictionary in your analysis, please verify assumptions about the data with the court.
Eviction case metadata (dataset #1)
The primary dataset is the Harris County Data Extract available from the Harris County Justice of the Peace Court. These data are available by querying up to 31 days of activity for a single court. This information comes from an application called the Justice Courts Management System (JCMS), which was built and maintained by the Harris County Justice Courts.
The full list of 46 fields include:
| Field | Description |
| Case Number | Unique identifier of the case in the dataset. |
| Case Type | The case type is CV to designate a civil case. |
| Case Subtype | The case subtype is Eviction. |
| Case File Date | The date the case was filed with the courts. Formatted %m/%d/%Y. |
| Style of Case | The names of the plaintiff and defendant, also known as a case caption. |
| Nature of Claim | The nature of claim is almost always Eviction, though there are a few rows of Debt Claim and Small Claims cases that are classified as an Eviction case subtype. Approximately 0.5% of cases have an NA in this field. |
| Claim Amount | Total amount of the original dispute. Approximately 2% of cases do not have a claim amount, and another 6.75% of cases have a claim amount of 0. |
| Case Status | There are five possible case statuses: Active, Appeal, Bankruptcy, Dismissed, and Disposed. The most common statuses are Dismissed and Disposed. Please note that the Case Status does not provide enough information to determine the outcome of the case. |
| Plaintiff Name | Name of the plaintiff. |
| Plaintiff Address | Address, city, state, and zip code of the plaintiff. |
| Plaintiff Attorney | Name of the plaintiff’s attorney. |
| Plaintiff Attorney Address | Address, city, state, and zip code of the plaintiff’s attorney. |
| Defendant Details | Defendant name and address. |
| Defendant Attorney Details | Defendant attorney’s name and address, if applicable. |
| Second Plaintiff Details | Name, address, city, state, and zip code of the second plaintiff, if applicable. |
| Second Plaintiff Attorney Details | Name, address, city, state, and zip code of the second plaintiff attorney, if applicable. |
| Second Defendant Details | Name, address, city, state, and zip code of the second defendant, if applicable. |
| Second Defendant Attorney Details | Name, address, city, state, and zip code of the second defendant attorney, if applicable. |
| Next Hearing Description | Generally, this will be populated with Eviction Docket, but it could also contain Default Docket, Dismissal Docket, Judge Trial, Jury Trial, Mediation, or Motion Docket. |
| Next Hearing Date | The date of the next upcoming hearing or the date of the most recent hearing in the case. Formatted %m/%d/%Y. |
| Next Hearing Time | The time of the next upcoming hearing or the date of the most recent hearing in the case. |
| Disposition Description | A short description of the disposition, such as Default Judgment, Dismissed for Want of Prosecution, Trial by Jury, Trial or Hearing by Judge, Non-suited or Dismissed by Plaintiff, Judgment Set Aside, Appeal Filed, and All Other Dispositions. |
| Disposition Date | The date of the disposition. Formatted %m/%d/%Y. |
| Judgment Text | A short description of the judgment, such as Default Judgment, Agreed Judgment, and Final Judgment. |
| Judgment Date | The date of the judgment. Formatted %m/%d/%Y. |
| Judgment In Favor Of | Name of the plaintiff or defendant who “won” the judgment. |
| Judgment Against | Name of the plaintiff or defendant who “lost” the judgment. |
| Judgment Amount | The final amount of the judgment that can be compared to Claim Amount. |
| Attorney Fees | Attorney fees that are awarded by the judge. |
| Court Costs | Court costs that are awarded by the judge. |
| Pre-Judgment Interest Rate | The interest rate charged on the claim amount before judgment. |
| Post-Judgment Interest Rate | The interest rate charged on the judgment amount after judgment. |
The preceding table is a modified version of the Harris County JP Court Civil File Specifications.
Enriched eviction case metadata (dataset #2)
In order to make the eviction case metadata easier to analyze, we transform and add several fields to the data file. If you are receiving an enriched data extract, rather than a raw data extract, it will include the following additional fields:
| Latitude | Geospatial coordinate derived from cleaning and geocoding the first defendant address. |
| Longitude | Geospatial coordinate derived from cleaning and geocoding the first defendant address. |
| JP court number | Justice of the Peace court and precinct number, derived from the initial data query. |
| Judge | Judge name derived from the presiding elected official at the time the case was filed. |
| Kinder Community Tabulation Area (KCTA) | Harris County “neighborhoods” assembled by the Kinder Institute at Rice University based on the building blocks of Census Tracts. This allows for additional analysis using Census/ACS data. |
| City councilmember | City council district where the eviction was located. |
| County commissioner | County commissioner precinct where the eviction was located. |
| GEOID | GEOID of the Census Tract where the eviction was located. |
| Census Tract name | Name of the Census Tract where the eviction was located. |
All of the geospatial elements, such as city council district, are derived using a point-in-polygon analysis from the coordinates of the defendant address.
Eviction docket scheduling data (dataset #3)
In addition to the eviction case metadata, we also collect cases that are listed on the individual judge’s docket. This is collected through a separate web scraper and cross-referenced against the eviction case metadata.
As part of this process, we also produce a dataset that includes:
| Field | Description |
| court_number | The precinct-place of the Justice court, such as 1-1 or 5-2. |
| docket_date | The date the case was scheduled on the eviction docket. Formatted %m/%d/%Y. |
| case_on_docket | The number of cases counted on the judges docket for that day, such as 23 or 55. |
| date_observed | The date the docket count was conducted. Formatted %Y-%m-%d. |
This dataset includes one row per observation, which means that there will be multiple rows for the same day/courtroom because we observed different case counts on the docket over time. This dataset can be useful to see how cases are moved onto and off of the docket as the hearing date approaches. It can also be useful to look at eviction docket peak times of year.
Eviction case events (dataset #4)
The eviction case events dataset is collected through the Harris County Odyssey case search portal. This requires querying an individual case and pulling up the series of associated events.
These eviction case events are useful for providing a timeline of the eviction case, as they include every action in the case that was filed with the court.
There are standard labels for the events (shown in blue), as well as free form comments (shown in gray). The event headers and comment sections were used extensively to mark policy interventions, such as the CDC eviction moratorium or the Texas eviction diversion program.
In order to assemble specific datasets for evaluating these policies, we collected case event data for all cases based on their last hearing date.
Because these cases are unfolding in real time, the eviction case event dataset should be seen as a snapshot in time. It should also be re-collected prior to any new analysis.
Eviction case documents (dataset #5)
The eviction case documents dataset is a selection of documents for 10,000 eviction cases that took place between 2018 and 2021.
The documents are scans of hardcopies that were filed with the court.
Because these documents are not standardized (and sometimes handwritten!), it is extremely difficult to assemble a large machine-readable dataset.
Instead, we often review a sample of eviction case documents to understand broad trends. For example, in a recent study of evictions after Hurricane Harvey, we found that 98% of evictions were filed due to nonpayment of rent.
Obtaining the data or providing feedback on the eviction data dictionary
Currently, the eviction case metadata is available through the Georgetown Civil Data Commons.
If you:
- Are a researcher working on an eviction-related project, and you would like access to more data
- Have questions about Harris County eviction data
- Have feedback about this eviction data dictionary
…please contact us!
