Skip to content

Data: Expungement Analysis

Jeremy Lang edited this page Jan 20, 2021 · 1 revision

Introduction

Our current data efforts for the MA clean slate project are towards answering the questions documented in this issue: https://github.com/codeforboston/clean-slate/issues/152

Primary Data sets

CrimeSOLV is crime data from police agencies in Massachusetts which report data to the Crime Reporting Unit in the format of the "National Incident Based Reporting System." The majority of police agencies now use this way of reporting their crime data. However, this is an incomplete data source (does not include all cities and towns, e.g. Boston) and cannot be considered state official statistics.

We can still run a preliminary analysis of this data.

Incident-based data by state, summary data with estimates, and data on specific topics like assaults on law enforcement officers, hate crime, or human trafficking are available for download in CSV files. Data is also available via the Crime Data API access to UCR data. The Uniform Crime Reporting (UCR) Program provided updated data for 2017 on September 24, 2018.

Other analyses / data work

Mapping of the master crime list to NIBRS incident categories

Analysis of the homogeneity of the FBI categories in the master crime list

Initial analysis of the master crime list by FBI category.

Cleaning and joining master crime list data to MA prosecution data

Massachusetts prosecution data cleaning (such as joining data and creating new columns) scripts can be found in the notebooks folder. Processed data in CSV format can be found in in the data/processed folder.

Currently, we are in the middle of identifying missing expungeability eligibilities for all of the unique charges that appear in the MA data from the Suffolk and Northwestern DA, and determining the minimal amount of information needed to deem a charge expungeable or not.

Other research topics

The above datasets are limiting in our estimates due to the following questions that we may not have explicit data for. However, we maybe able to apply some simple overlays on the numbers we get above to arrive at a more precise estimate. The following are sorted in a manner to prioritize items that have a higher impact to the overall counts:

  • How many marijuana related arrests were there that are now expungeable?
  • What are the number or percentage of crimes committed concurrently / number of people who have uniquely committed a crime?
  • What percent of arrests result in dropped charges?
  • What are the actual conviction rates for different types of arrests? charges?
  • What percent of crimes involve a motor vehicle offense?
  • What percentage of crimes within FBI NIBRS categories contain crimes that are expungeable and some that are not?
  • How long does it take from arrest to conviction?
  • How many of the arrests/charges/convictions were for crimes that are now legal excluding marijuana?
  • What percent of arrests/charges result in mistrial / dropped charges / plea deal?

Open Tasks

All open tasks have been migrated to Data Branch Issue Tracking Project. Check there for an up-to-date view of current priorities.

Clone this wiki locally