OpenDataPortalforChampaignRacialandCriminalJustice

TowardsGreaterTransparencyinPolicyMaking

YuyanHuangMasterofUrbanPlanning,2017UniversityofIllinoisatUrbana-ChampaignAdvisor:Prof.BevWilson

Abstract

This project focuses on the application of civic technology to urban planningpractice in racial justice. It explores the capability of computer technology toimprovetheaccessibilityofcivicdataandtobringgreatertransparencytopublicdecision-making.WorkingincollaborationwithChampaignCountyRacialJusticeTaskForce,theauthordevelopedawebsite1 anditsbehinddatabasetosupportthevisualizationofpubliclyavailableinformationonarrests,correctionalfacilityrecordsandcourtcasesinUrbana-Champaign.

ThewebsiteanddatabasehasreceivedpositivefeedbackfromtheTaskForceasit collects and centralizes data scattered in Champaign criminal justice systemand responds to data needs brought forward by civil society and localgovernmentregardinglawenforcement.Withthistool,communities,interestedsocialgroupsandindividualswillhaveeasieraccesstolegaldataandbebetterinformedofpossibleracialdisparitiesinthecriminaljusticesystem.

ProjectBackground

RacialdisparityinthecriminaljusticesystemhaslongbeenavexingproblemintheUnitedStates.EventsofpoliceshootingofunarmedAfrican-Americanshavebroughtthisissuetothespotlightandarousedpeople’squestioningoftheethicsof policepractice and thedisparity toward communities of color. Studies haveshownthatsubtlediscriminationexists in thecriminal justicesystematnearlyevery stage from arrest rates, bail amounts, to sentence length and probationhearingoutcomes.

UrbanPlanningisadisciplinededicatedtoenhancesocialequity.Exploringdatain the criminal justice system and understand their implications are of greatimportanceforplannersandpolicy-makerstoaddressracialdisparityproblemsandtomakebetterandmoretransparentpolicydecisions.

However, the lack of access to well-formatted and readable data in criminaljusticesystemhascreatedobstaclesinthisprocess.InChampaignCounty,dataonpolicerecords,jailpopulation,bondinformationandcourtcasesarereleasedbydifferentagenciesindifferentdataformats.Forexample,policearrestrecordsof Champaign City, Urbana City and the University are published on threedifferent websites as spreadsheets with different attributes. Information oninmateswhoareheldinoneofthecounty’scorrectionalfacilitiesarescatteredindaily jail log records, jail reports and an inmate lookup website. Without acomprehensive and effective database and an open data portal, it would bechallenging for individuals, local groups and organizations to access and usethesedatatofacilitatefurtherresearch.

1 Linktothewebsite:https://racialjustice.herokuapp.com/

Thisproject aims tobuildadatabaseandawebsite to centralizeandvisualizedatascatteredinChampaigncriminaljusticesystem.ItisbuiltonthefoundationofapreviousprojectcollaborativelyfinishedbyfifteenUIUCstudentstakingthecourseCivictechnologyandDigitalCity.Studentsaredividedintothreegroups–police,correctionalfacilityandcourt.Eachgroupwasresponsibleforcollectingdataincorrespondingfieldandbuildingvisualizationdashboardusingshinyappframework in R. At the end of the course, Prof. Wilson incorporated threedashboardsintoafinalone2 andpublisheditusingashinyserverhostedontheROGERsupercomputeratUIUC.

Figure1.ASnapshotofthe“Resolution”PageofPreviousPoliceDashboard3

Tomakethesedashboardsmorepubliclyaccessibleandtomaintainalong-termonlinedatabase,theauthorbuiltamorecomprehensivewebsiteanditsbehinddatabase,whichare currentlyhostedon theHeroku server.Visualization toolsthatareoriginallywritteninRareallrewrittenusingtheDjangowebframeworkinvolving front-endprogramming languages, suchasHTML,CSSand JavaScriptand back-end programming languages, such as Python and PostgreSQL. Userscannowvisitthewebsiteathttps://racialjustice.herokuapp.com/.AllcodeshavebeenpublishedonGitHub:https://github.com/yhung133/racial-justice-app.

2 Linktothefinaldashboard:http://141.142.170.106/VizTools/3 Linktothepolicedashboard:https://yhung133.shinyapps.io/Project_Final/

EndProductOverview

Thiswebsiteisdesignedtohavefourmajorcomponents:

Figure2.LandingPageoftheWebsiteShowingtheFourComponents

1. DataStories

Previousvisualization toolsaremainlydescriptivechartsor tablesof collecteddata.However,totrulyaddressracialdisparities inthecriminal justicesystem,deeper statistical analyses areneeded. In collaborationwith students from theDepartmentofStatisticsatUIUCwhoaretakingthecourseStatisticalConsulting,thispartof thewebsite isdesigned tohighlightstatisticallysignificant findingsregarding racial disparities in a story-telling way. Students from the STATDepartmentareresponsibleforanalyzingcollectedarrest, jailpopulation,bondpayment,andcourtcasedatausing theirknowledge instatistics.Theauthor isresponsible for visualizing their findings and embedding these results into thewebsiteplatform.

2. DataVisualization

Figure3.SummarizationWebpageofThreeDashboards

Datavisualizationpartofthewebsiteisdesignedtovisualizedatasetsstoredintheonlinedatabaseusinginteractivebarcharts,linecharts,piechartsandtables.Mostvisualizationtoolsare“translated” frompreviousdashboards.Theyallowuserstoexploredescriptiveillustrationofcriminaljusticedataandhaveabasicunderstandingoftrendsandcharacteristics.

This part is further composed of three webpages corresponding to threeprevious dashboards (policing, correctional facilities, and courts), which areindividually finishedby three student groups taking the courseCivicTech andDigitalCity.

1)Policing

The Policing dashboard focuses on visualizing characteristics of arrestedpopulation. Collected policing data include incidents, arrests and reportedcrimes fromeachof the fourpoliceagencies thatparticipate in thecentralizedpolicedatamanagementsystemincluding:CityofChampaign,CityofUrbana,UofIPolice,andtheChampaignCountySheriff.Thesedatawereprovidedbytheselocal police agencies through a formal FOIA request. Arrests include policeincidents in which an individual is charged with a violation of the law. Thismeans the arrest data includes data for individualswhowere issued citationsandtraffic tickets,alongwith individualswhowerechargedwithmoreseriouscrimes.

2)CorrectionalFacilities

This dashboard visualizes data about in-jail population. The correctionalfacilitiesworkinggroupcollecteddemographicinformationaswellaslegaldata(e.g.bonds,charges,courtdates,releasedates,etc.)oninmateswhowereheldin

oneofthecounty’sfacilitiesbeginningJanuary1,2010untilSeptemberof2016(i.e.,thedateoftherequests).Theyvisualizetheracialbreakdownofpopulationin correctional facilities in Champaign County and provide real-time numbersand interactive charts that allows users to have easier access to legal data oninmateswhowereheldinoneofthecountyfacilitiesandtogetmoreinsightsonracialdisparityincriminaljusticesystem.

Figure4.SnapshotofPartofThePolicingDashboardWebpage

3)Court

The court groupworkson court case records fromCircuitClerkofChampaign

County.SincetheFOIArequestaskingforrecordsprovidedintheclerk’sonlinepublic access system hasn’t been responded, this group developed a series ofscripts to 'scrape' thedesired information fromtheCircuitClerkofChampaignCountywebsite as a temporarywork-around in hopes that the requesteddatawill be provided at a future date. The court group visualizes disparities inchargesandsentencing.

3. DataPortal

The Data Portal part is designed to show a list of data tables archived in thedatabase. Each table comes with a brief description of its source and specificattributes.Thispartallowsuserstoviewformatteddatatablesthatsupportthevisualizationtoolsanddownloadthesedatasetsforfurtherresearchuse.

4. Community

Localcommunityengagementisanessentialpartintheimplementationofurbanplanningpractices.Quantitativedata,suchaschartsandtables,arenotenoughtoarouseattentionofpotentialracialdisparitiesinlocalcriminaljusticesystemand to promote corresponding public policy making. To supplement collectedquantitative data, the Champaign Racial Justice Task Force proposed to haveanother database collecting and archiving qualitative data, such as videos ofinterviews with community members sharing their stories about racialdisparities,especiallythoserelatedtocriminaljustice.

This part is alsodesigned to allowwebsite users to register their account andmakerelevantpostingsontheonlinediscussionforum.

Database,DataDescriptionandSources

Adatabaseisacollectionofinformationthatisorganizedsothatitcanbeeasilyaccessed,managed and updated4. There are generally two types of database –relational database and non-relational database. A relational database is acollection of data items organized as a set of formally described tables fromwhichdatacanbeaccessedore reassembledwithout changing thepre-definedschema of tables5. The programming language to interact with a relationaldatabaseisthestructuredquerylanguage(SQL).

There are two major advantages of storing data tables into a SQL databasecomparedtojustsavingtheminExcelspreadsheetorCSVformat:1)tablesarebetterformattedsoitcanbeeasilyexpandedandmaintainedinthelongterm;2)data stored in a relational database can be quickly accessedwhendrawing aninteractivechartsusingtheSQLprogramminglanguage.

4 Whatisdatabase(DB)?-DefinitionfromWhatIs.com.(n.d.).RetrievedAugust04,2017,fromhttp://searchsqlserver.techtarget.com/definition/database5 Whatisrelationaldatabase?-DefinitionfromWhatIs.com.(n.d.).RetrievedAugust04,2017,fromhttp://searchsqlserver.techtarget.com/definition/relational-database

Previously,whenwewerebuildingthedashboardsusingShinyAppframeworkinR,all tablesweresaved inaSQLitedatabase.SQLite isan in-process librarythat implements a self-contained, server-less, zero-configuration, andtransactional SQL database engine6. However, SQLite is not comparable toclient/serverSQLdatabaseenginessuchasMySQL,Oracle,PostgreSQL,orSQLServer. Client/server SQL database engines strive to implement a sharedrepository of enterprise data. They emphasize scalability, concurrency,centralization, and control. SQLite strives to provide local data storage forindividual applications and devices. SQLite emphasizes economy, efficiency,reliability,independence,andsimplicity7.

Simply speaking, SQLite ismore like a local contained storing structured datatables. But when it comes to support a website, it cannot compete withclient/server databases. Therefore, the author switched to a PostgreSQLdatabasewhenbuildingthewebsite.

Figure5.ASnapshotofPreviousSQLiteDatabase

Six tables are currently stored in the PostgreSQL database, with one tablecovering police arrest records, four tables covering in-jail populationcharacteristicsaswellasbondpaymentinformation,andanothertablecoveringcourt case records. Some of these tables come from FOIA (Freedom ofInformation Act) requests. Others are scraped from publicly accessiblegovernmentwebsites.Moredetailedinformationabouteachtableisshowedasfollows:

TableInformation

TableAttributes TableSource

6 AboutSQLite.(n.d.).RetrievedAugust04,2017,fromhttps://www.sqlite.org/about.html7 AppropriateUsesForSQLite.(n.d.).RetrievedAugust04,2017,fromhttps://sqlite.org/whentouse.html

Police

ArrestRecords

ID,DateofArrest,TimeofArrest,ArrestCode,LocationofArrest,CrimeCode,CrimeCodeDescription,CrimeCodeCategoryWeaponCode1,WeaponCode2,WeaponCode3,RaceDescription,SexDescription,AgeatArrest,EmploymentCode,ArrestingOfficerBadgeNumber,Result,PoliceDepartment,Morethanonecharge,MVO(dummyvariableifoneofthechargesisMotorVehicleOffence)

FOIARequestSubmittedon11/17/2016

Jail

Inmate

LookUp

Name,BookedData,BookedTime,Age,Race,Address,Bond,Charges,CourtType,CourtDate

ChampaignCountySherifOfficeInmateLookupSystemhttp://www1.co.champaign.il.us/SHERIFF/InmateLookUp.php

InmateRecords

BookedDate,BookedTime,DOB,Age,Gender,Race,(Number,Direction,Street,Suffix,City,State,ZipCode,Statue#),Description,CourtDate,ReleaseDate,BondAmount

FOIARequestSubmittedon11/17/2016

DailyJailLog Name,Race,Age,Sex,Agency,BookDate,BookTime,JacketNo.,ReportNo.,Charges,(BookNRelease,ReleaseonRecognizance(PersonalorCCSO),Bondpaid,ServedSentence,Release,EntryNo.,NumberofCharges),(CityOrdinanceViolation,DrivingonSuspendedorRevokedLicense,FailuretoAppear,ResistorObstructingaPeaceOfficer,DomesticBattery,OtherBattery,Charges)

OriginallyreleasedbyChampaignCountySherifOfficeWebsitehttp://www.co.champaign.il.us/Sheriff/DailyReports.phpCollectedandprocessedbyNCSAatUIUChttp://dart.ncsa.uiuc.edu/stuffed/bpnj/daily_jail_log/

DailyJailArraignment

Name,Jacket#,CourtDocketNumber,ReportNumber,BookedDate,BookedTime,Charges

ReleasedbyChampaignCountySherifOfficeWebsitehttp://www.co.champaign.il.us/Sherif

f/DailyReports.php

Court

CourtCaseRecords

CaseNumber,FileDate,Race,Sex,Age,ChargeType(DUIclassA,FelonyClass1,FelonyClass2,FelonyClass3,FelonyClass4,… MisdemeanorClass4,MisdemeanorClassA,…),ChargeAgency,Charges,DispositionType,Defendant.Plea,SentenceDate,Sentence.

ReleasedbyChampaignCircuitClerkWebsitehttp://www.cccircuitclerk.com/

Table1.SummarizationofCollectedDataTables

TechnicalProcedure

1)Front-endProgramming–HTML,CSSandJavaScript

The“front-end”ofawebsitesimplymeansthepartofawebsiteausercanseeandinteractwith.EverywebpageauserseeisacombinationofHTML,CSSandJavaScript,whichareallbeingcontrolledbythewebbrowser,suchasChromeorSafari,oftheuser’scomputer

HTML(HypertextMarkupLanguage)isthemostbasicbuildingblockoftheWeb.Itdefines thecontentofawebpage8.CSS isused tostylize theappearanceandpresentationofHTMLelements.JavaScriptisusedtoaddfunctionalitytoHTMLelements,makingawebpageinteractive.

Forthisproject,eachwebpageiswritteninoneHTMLdocument.MostoftheCSSstylizedelementsare inherited fromBootstrap–apopular frameworkused tobuildresponsive,mobile-firstprojectsontheweb9.

2)Back-endProgramming–Python,DjangoWebFramework,PostgreSQL

The back end of awebsite consists of a server, an application and a database,whichsupportthefunctionalityofthefront-end.Forthisproject,theauthorusedthe Django Web Framework to develop the whole application. Django is ahigh-levelPythonWebframeworkthatencouragesrapiddevelopmentandclean,pragmaticdesign10.As for thedatabase,PostgreSQL is selectedbecause it isafreeadd-onprovidedbyHeroku.

3)DeploymentonHeroku

Herokuisa freecloudplatformthat letsdevelopersbuild,deliver,monitorand8 HTML.(n.d.).RetrievedAugust04,2017,fromhttps://developer.mozilla.org/en-US/docs/Web/HTML9 MarkOtto,JacobThornton,andBootstrapcontributors.(n.d.).Bootstrap.RetrievedAugust04,2017,fromhttps://v4-alpha.getbootstrap.com/10 Django.(n.d.).RetrievedAugust04,2017,fromhttps://www.djangoproject.com/

scale apps11.Herokuhelpsdevelopers saving tonsof timeneeded todealwithhundredsofcloudinfrastructuredeploymentscriptsorrackingserverstoscale.Besides,Herokuprovides secureandscalabledatabase, suchasPostgreSQL, tosupportitshostedapplication.

Figure6.ASimpleIllustrationoftheWholeWebDevelopmentFramework

FutureWork

1. FurtherCollaborationwithSTATStudents

Asismentionedabove,totrulyaddressracialdisparitiesinChampaigncriminaljusticesystem,statisticalanalysesofcollecteddataarenecessary.StudenttakingthecourseStatisticalConsultingweretryingtoanswerthefollowingquestions:

• Whatistheproportionofarresteesunderdifferentdemographicfactorscomparedtocensusdata?

• Given that someone is already arrested, is race a significant factor thatwould influencetheoutcomeofwhetheranarresteewouldbetakentojail?

• Arethewaitingtime,releasetime,andbondamountdifferentforpeoplewhoarebookedduetorace?

11 WhatisHeroku?(n.d.).RetrievedAugust04,2017,fromhttps://www.heroku.com/what

• Are African Americans charged with felonies more likely to beimprisoned?

• Is there a significant difference between the lengths of jail time due torace?

However,duetothelimitationofcollecteddata,mostoftheanswersturnouttobestatistically insignificant, indicatingtherearenosignificantracialdisparitiesinthecriminaljusticesystem,whichisincontradictorywiththesomefactsweobserved. To solve these contradictories, we need to collect better-formatteddataanddoingmorestatisticalanalyses.

2. Databaseexpansionandmaintenance

Currently the database can only be periodically updatedmanually since somedata can only be collected fromFOIA requests, ofwhich the response and thetimeofresponsecannotbeguaranteed.Inthefurther,tofacilitatetheautomaticupdate of the database, the author has to further collaborate with the RacialJusticeTaskForcetofindoutamorestandardproceduretocollectrelevantdata.

3. RefinementofDescriptiveVisualizationTools

Some of the current visualization tools need to be refined to have a clearerpresentationofdescriptivecharacteristicsofcollecteddata.Forexample,inonechartofthejaildashboard,theracialcompositionofdifferentresults(including:BailPaid”, “BookNRelease”, “ReleaseonRecognizance”and “Still in Jail at theEndDayofBooking”) for inmates at the endof thebookingday ismisleading.This chart didn’t specify crime types the inmates have conducted. However,resultsareheavily influencedby thecrimes types insteadof just influencedbyraces.Toget ridof thesemisleadingcharts, suchvisualization toolshave toberefined.

4. DevelopmentoftheVideoPlatform

Localcommunityengagementiscriticalintheenforcementofplanningpractices.To arouse the local community’s attention of potential racial disparity in thecriminal justicesystem,CarolynRandolph,amemberof theRacial JusticeTaskForce,proposedtohaveanotherdatabasearchivingvideointerviewswithlocalresidentswhohave experienced racial disparities in the legal process. CarolynRandolph, as a PHD in sociologywould lead a team to prepare and finish therecordingofvideointerviews,whiletheauthorwouldberesponsibletobuildthevideodatabaseandpresentvideosonthewebsite.

References

Sampson,R.J.,&Lauritsen,J.L.(1997).RacialandethnicdisparitiesincrimeandcriminaljusticeintheUnitedStates.Crimeandjustice,21,311-374.

Harris, D. A. (2003). The reality of racial disparity in criminal justice: Thesignificanceofdatacollection.LawandContemporaryProblems,66(3),71-98.

Crutchfield,R.D.,Fernandes,A.,&Martinez,J.(2010).Racialandethnicdisparityandcriminal justice:Howmuch is toomuch?.The JournalofCriminalLawandCriminology(1973-),100(3),903-932.

Pettit,B.,&Western,B.(2004).Massimprisonmentandthelifecourse:Raceandclass inequality in US incarceration. American sociological review, 69(2),151-169.

Whatisdatabase(DB)?-DefinitionfromWhatIs.com.(n.d.).RetrievedAugust04,2017,fromhttp://searchsqlserver.techtarget.com/definition/database

What is relational database? - Definition from WhatIs.com. (n.d.). RetrievedAugust04,2017,from

http://searchsqlserver.techtarget.com/definition/relational-database

AboutSQLite.(n.d.).RetrievedAugust04,2017,from

https://www.sqlite.org/about.html

AppropriateUsesForSQLite.(n.d.).RetrievedAugust04,2017,from

https://sqlite.org/whentouse.html

HTML.(n.d.).RetrievedAugust04,2017,fromhttps://developer.mozilla.org/en-US/docs/Web/HTML

MarkOtto,JacobThornton,andBootstrapcontributors.(n.d.).Bootstrap.RetrievedAugust04,2017,fromhttps://v4-alpha.getbootstrap.com/

Django.(n.d.).RetrievedAugust04,2017,fromhttps://www.djangoproject.com/

WhatisHeroku?(n.d.).RetrievedAugust04,2017,fromhttps://www.heroku.com/what

Appendices

Figure7.SnapshotoftheProject’sGitHubRepository

Figure8.SnapshotoftheProject’sDevelopmentEnvironment