MCP Server Proto-OKN

# SCALES Knowledge Graph Overview and Class Diagram ## Chat Transcript ### Introduction to SCALES **Date:** December 15, 2025 **Overview:** SCALES (Systematic Content Analysis of Litigation Events) is an open knowledge network designed to democratize access to criminal and civil court records across the United States. --- ## SCALES Knowledge Graph Description ### Project Information - **Endpoint:** https://frink.apps.renci.org/scales/sparql - **Homepage:** https://scales-okn.org/ - **Funding:** NSF Award #2333803 - **Statistics:** https://frink.renci.org/kg-stats/scales-kg ### Principal Investigators - **Danny O'Neal** - danny.e.oneal@gmail.com - **Scott Daniel** - scottgdaniel@gmail.com ### Knowledge Graph Statistics The SCALES knowledge graph contains: - **523+ million triples** describing relationships between entities - **96.5 million entities** across the justice system - **4.1+ million cases** total - 2.4 million criminal cases - 1.8 million civil cases - Data from **federal district courts** and **state/local court systems** ### Data Coverage The knowledge graph integrates and links: - **Parties** (plaintiffs, defendants, organizations) - **Attorneys** (defense, prosecution, civil representatives) - **Law Firms** - **Judges** - **Charges** (criminal and civil) - **Sentences** (imprisonment, probation, fines) - **Arrests** - **Bookings** (jail intakes) - **Courts** (federal, state, local) - **Docket Entries** (case actions and filings) ### Standards and Purpose Built on the **National Information Exchange Model (NIEM) Justice Domain** standard, SCALES employs standardized vocabularies to: - Enable systematic analysis of justice system functioning - Reveal inequities in the criminal justice system - Support evidence-based policy making - Address the critical lack of nationally-accessible and linked criminal justice data --- ## Schema Overview ### Main Entity Classes (24 total) #### Case Types 1. **CriminalCase** - Criminal court cases 2. **CivilCase** - Civil court cases #### People 3. **Person** - Individual people in the system 4. **Judge** - Judicial officials 5. **Attorney** - Legal counsel 6. **CaseDefendantParty** - Defendants in cases 7. **CaseInitiatingParty** - Plaintiffs/prosecutors 8. **CaseDefenseAttorney** - Defense attorneys on cases 9. **CaseInitiatingAttorney** - Prosecuting/plaintiff attorneys 10. **CaseJudge** - Judges assigned to cases 11. **LawEnforcementOfficial** - Police and law enforcement #### Organizations 12. **Firm** - Law firms 13. **Party** - Legal parties (persons or organizations) 14. **Court** - Courthouses and judicial venues 15. **BookingFacility** - Jails and detention facilities #### Legal Entities 16. **Charge** - Criminal or civil charges 17. **ArrestCharge** - Charges at time of arrest 18. **Sentence** - Court-ordered sentences 19. **SentenceTerm** - Components of sentences 20. **Arrest** - Arrest events 21. **Booking** - Jail booking/intake events 22. **Release** - Release from custody events #### Documents 23. **RegisterOfActions** - Docket tables 24. **RegisterAction** - Individual docket entries ### Key Properties (259 total predicates) The knowledge graph includes properties for: #### Case Information - Docket IDs and case numbers - Filing dates, start dates, end dates - Case status and disposition - Court circuits and jurisdictions - Case categories and types #### Charge Details - Offense codes (NIBRS, UCR) - Charge text and descriptions - Severity levels (felony, misdemeanor) - Disposition outcomes - Statute references #### Sentencing Information - Prison terms and duration - Probation periods - Fine amounts - Restitution orders - Supervised release terms #### Person Demographics - Full names and IDs - Race and ethnicity - Sex/gender - Birth year - Residence information #### Geographic Data - Court locations and postal codes - County codes (FIPS) - Census block IDs - Facility addresses #### Professional Relationships - Attorney-client assignments - Judge assignments to cases - Firm affiliations - Party representations --- ## Example Data ### Sample Criminal Case - **Case ID:** akd;;1:16-cr-00001 - **Docket:** 1:16-cr-00001 - **District:** akd (Alaska District) - **Circuit:** 9.0 (Ninth Circuit) - **Filed:** 2016-02-03 - **Closed:** 2016-08-04 - **Status:** closed ### Sample Charges 1. **26:7203 WILLFUL FAILURE TO FILE INCOME TAX RETURNS** - Counts: 1-3 - Disposition: 12 months 1 day imprisonment; 1 year Supervised Release; $600,064.00 Restitution; $100.00 Special Assessment 2. **18:2252(a)(2) and (b)** - Count: 1 - Disposition: 60 months imprisonment; 20 years supervised release; $100 special assessment; property forfeiture 3. **18:1159(a) AND (b)(1)(B)(I) MISREPRESENTATION OF INDIAN PRODUCED GOODS** - Counts: 1-2 - Disposition: 5 years Probation; $4000.00 Fine; $4000.00 Restitution; $50.00 Special Assessment ### Sample Person Data Persons include demographic information: - **Sex Code:** M/F - **Race Text:** B (Black), W (White), etc. - **Birth Year:** e.g., 1975, 1992, 1980 --- ## SCALES Knowledge Graph Class Diagram ```mermaid classDiagram %% Core Case Classes class CriminalCase { +CaseDocketID: string +CaseGeneralCategoryText: string +StartDate: date +EndDate: date +StatusDescriptionText: string +hasIdbCircuit: float +hasIdbCounty: float } class CivilCase { +CaseDocketID: string +CaseGeneralCategoryText: string +CaseSubCategoryText: string +StartDate: date +EndDate: date } %% Legal Entities class Charge { +ChargeText: string +ChargeSequenceID: string +ChargeDisposition: string +ChargeDispositionCategoryText: string +ChargeSeverityLevelCode: string +hasChargeType: string } class Sentence { +SentenceDescriptionText: string } class SentenceTerm { +TermDuration: string } class Arrest { +ActivityDate: date } class ArrestCharge { +ChargeText: string } class Booking { +ActivityDate: date } class Release { +ActivityDate: date +BondAmount: decimal +BondType: string } %% People and Organizations class Person { +PersonFullName: string +PersonID: string +PersonRaceText: string +PersonSexText: string +PersonSexCode: string +BirthYear: string } class Attorney { +PersonFullName: string +OrganizationName: string } class Judge { +PersonFullName: string +JudicialOfficialCategoryText: string +hasUVAJudgeDirID: string +hasCommissionDate: date } class CaseDefendantParty { +ParticipantRoleCategoryText: string } class CaseInitiatingParty { +ParticipantRoleCategoryText: string } class CaseDefenseAttorney { +CaseOfficialRoleText: string } class CaseInitiatingAttorney { +CaseOfficialRoleText: string } class CaseJudge { +CaseOfficialRoleText: string } class Firm { +EntityName: string +OrganizationName: string } class Party { +PersonFullName: string +EntityName: string } class LawEnforcementOfficial { +PersonFullName: string } %% Locations and Courts class Court { +CourtName: string +CourtCategoryCode: string +CourtPostalCode: string +JurisdictionText: string +isInCircuit: string } class BookingFacility { +FacilityName: string +AddressPostalCode: string } %% Documents class RegisterOfActions { +AdministrativeID: string } class RegisterAction { +RegisterActionDate: date +RegisterActionDescriptionText: string } %% Relationships - Criminal Case CriminalCase "1" --> "0..*" Charge : CaseCharge CriminalCase "1" --> "1" Court : CaseCourt CriminalCase "1" --> "0..*" CaseDefendantParty : CaseDefendantParty CriminalCase "1" --> "0..*" CaseInitiatingParty : CaseInitiatingParty CriminalCase "1" --> "0..*" CaseJudge : CaseJudge CriminalCase "1" --> "0..1" RegisterOfActions : DocketTable %% Relationships - Civil Case CivilCase "1" --> "1" Court : CaseCourt CivilCase "1" --> "0..*" Party : CaseInitiatingParty CivilCase "1" --> "0..*" Party : CaseDefendantParty CivilCase "1" --> "0..*" CaseJudge : CaseJudge %% Relationships - Charge Charge "0..*" --> "0..1" Sentence : ChargeSentence Charge "0..*" --> "1" Person : ChargeSubject Charge "0..*" --> "0..1" Booking : ChargeBooking %% Relationships - Sentence Sentence "1" --> "0..*" SentenceTerm : SentenceTerm %% Relationships - Arrest and Booking Arrest "1" --> "1" Person : ChargeSubject Arrest "1" --> "0..*" ArrestCharge : PersonCharge Arrest "1" --> "0..1" LawEnforcementOfficial : ArrestOfficial Arrest "1" --> "0..1" Booking : SubjectBooking Booking "1" --> "1" BookingFacility : BookingFacility Booking "1" --> "0..1" Release : BookingRelease Booking "0..*" --> "0..*" CriminalCase : BookingCase %% Relationships - People Roles CaseDefendantParty "1" --> "1" Person : isInstanceOfEntity CaseInitiatingParty "1" --> "1" Person : isInstanceOfEntity CaseDefenseAttorney "1" --> "1" Attorney : Attorney CaseDefenseAttorney "0..*" --> "0..1" CaseDefendantParty : assignedToDefendant CaseInitiatingAttorney "1" --> "1" Attorney : Attorney CaseInitiatingAttorney "0..*" --> "0..1" Party : appointedByParty CaseJudge "1" --> "1" Judge : isInstanceOfEntity %% Relationships - Organizations Attorney "0..*" --> "0..1" Firm : Firm %% Relationships - Documents RegisterOfActions "1" --> "0..*" RegisterAction : DocketEntry RegisterAction "0..*" --> "0..*" RegisterAction : hasReferenceToOtherEntry ``` --- ## Class Diagram Explanation ### Entity Relationships #### Criminal Case Flow 1. A **CriminalCase** has one or more **Charges** 2. Each **Charge** is associated with a **Person** (defendant) 3. **Charges** may lead to **Sentences**, which have **SentenceTerms** 4. Cases are heard in a **Court** by a **CaseJudge** 5. **CaseDefendantParty** and **CaseDefenseAttorney** represent the defense 6. **CaseInitiatingParty** and **CaseInitiatingAttorney** represent the prosecution #### Arrest and Booking Flow 1. An **Arrest** involves a **Person** and a **LawEnforcementOfficial** 2. **Arrests** have **ArrestCharges** 3. An **Arrest** may lead to a **Booking** at a **BookingFacility** 4. **Bookings** may result in a **Release** (with bond information) 5. **Bookings** are linked to **CriminalCases** #### Civil Case Flow 1. A **CivilCase** involves multiple **Parties** (initiating and defendant) 2. **Parties** can be represented by **Attorneys** from **Firms** 3. Cases are presided over by **Judges** via **CaseJudge** role 4. Cases are heard in specific **Courts** #### Document Trail 1. Cases have a **RegisterOfActions** (docket table) 2. Each **RegisterAction** is a docket entry (filing, hearing, order, etc.) 3. **RegisterActions** can reference other entries ### Cardinality Notes - **1** = Exactly one - **0..1** = Zero or one (optional) - **0..\*** = Zero or more (many) For example: - A CriminalCase has exactly **1** Court - A CriminalCase has **0 or more** Charges - A CriminalCase has **0 or 1** RegisterOfActions (docket table) --- ## Use Cases ### Research Applications 1. **Sentencing Disparities Analysis** - Compare sentences across demographics (race, gender, age) - Analyze geographic variations in sentencing - Identify patterns in charge disposition 2. **Attorney Performance Tracking** - Analyze case outcomes by attorney - Study public defender vs. private attorney effectiveness - Examine law firm specializations 3. **Judicial Behavior Studies** - Track judge assignment patterns - Analyze sentencing tendencies by judge - Study case load distributions 4. **Criminal Justice System Flow** - Map arrest-to-sentencing timelines - Analyze booking and release patterns - Study bail and bond decisions 5. **Geographic Justice Patterns** - Compare outcomes across districts and circuits - Analyze urban vs. rural justice administration - Study county-level variations 6. **Charge and Offense Analysis** - Track offense trends over time - Analyze charge reduction patterns - Study plea bargain outcomes --- ## SPARQL Query Examples ### Example 1: Get Cases in a Specific District ```sparql PREFIX scales: <http://schemas.scales-okn.org/rdf/scales#> PREFIX nc: <http://release.niem.gov/niem/niem-core/5.0/> PREFIX jxdm: <http://release.niem.gov/niem/domains/jxdm/7.2/> SELECT ?case ?docketID ?startDate ?status WHERE { ?case a scales:CriminalCase . ?case nc:CaseDocketID ?docketID . ?case nc:StartDate ?startDate . ?case nc:StatusDescriptionText ?status . ?case scales:hasIdbDistrict "akd" . } LIMIT 10 ``` ### Example 2: Get Charges and Sentences ```sparql PREFIX jxdm: <http://release.niem.gov/niem/domains/jxdm/7.2/> SELECT ?charge ?chargeText ?disposition WHERE { ?charge a jxdm:Charge . ?charge jxdm:ChargeText ?chargeText . ?charge jxdm:ChargeDisposition ?disposition . } LIMIT 10 ``` ### Example 3: Get Person Demographics ```sparql PREFIX nc: <http://release.niem.gov/niem/niem-core/5.0/> PREFIX jxdm: <http://release.niem.gov/niem/domains/jxdm/7.2/> PREFIX scales: <http://schemas.scales-okn.org/rdf/scales#> SELECT ?person ?race ?sex ?birthYear WHERE { ?person a nc:Person . ?person nc:PersonRaceText ?race . ?person jxdm:PersonSexCode ?sex . ?person scales:BirthYear ?birthYear . } LIMIT 10 ``` --- ## Technical Details ### Data Standards - **NIEM (National Information Exchange Model)** - Version 7.2 Justice Domain - **NIEM Core** - Version 5.0 - **FBI NIBRS** - National Incident-Based Reporting System codes (2023.0) - **FIPS Codes** - Federal Information Processing Standards for counties ### Ontology URIs - **JXDM Namespace:** `http://release.niem.gov/niem/domains/jxdm/7.2/` - **NIEM Core Namespace:** `http://release.niem.gov/niem/niem-core/5.0/` - **SCALES Namespace:** `http://schemas.scales-okn.org/rdf/scales#` - **FBI NIBRS Namespace:** `http://fbi.gov/cjis/nibrs/2023.0/` ### Access Points - **SPARQL Endpoint:** https://frink.apps.renci.org/scales/sparql - **Triple Pattern Fragments:** https://frink.apps.renci.org/ldf/scales - **Statistics Dashboard:** https://frink.renci.org/kg-stats/scales-kg --- ## Contact Information For questions, contributions, or collaborations: - **Danny O'Neal** - danny.e.oneal@gmail.com (GitHub: @danny-oneal) - **Scott Daniel** - scottgdaniel@gmail.com (GitHub: @scottgdaniel) --- ## Additional Resources - **Project Homepage:** https://scales-okn.org/ - **NSF Award:** https://www.nsf.gov/awardsearch/showAward?AWD_ID=2333803 - **Documentation:** Available in the SCALES repository --- ## Summary The SCALES knowledge graph represents one of the most comprehensive, linked datasets of criminal and civil justice records in the United States. By standardizing data across jurisdictions and connecting entities like cases, charges, people, attorneys, judges, and courts, SCALES enables researchers, policymakers, and advocates to: - Conduct large-scale empirical studies of the justice system - Identify and measure disparities in outcomes - Track the flow of cases through the system - Analyze attorney and judicial behavior - Support evidence-based criminal justice reform With over 523 million triples covering millions of cases, SCALES provides unprecedented visibility into the functioning of the American justice system. --- *Document Generated: December 15, 2025* *SCALES Knowledge Graph - Systematic Content Analysis of Litigation Events*