Evidence database for law enforcement – Initial PHP/MySQL version

Purpose

This project was developed to aggregate and centralize evidence of harassment and stalking incidents, providing fast, user-friendly, searchable access for law enforcement and the district attorney handling the case. It enabled investigators and attorneys to efficiently filter and review various forms of digital contact between the harasser and victims.

Project turnaround was of utmost importance as it dealt with an ongoing case. The original version of this project was created in about a day and a half with a rewrite a few months later.

I have been unable to find a copy of the original PHP version of this project or I would provide a GitHub link. However, the code for the initial PHP-based tools I wrote to import emails/texts from different sources can be found here.

Tech Stack

• Docker – provides containerization for MySQL and both NGINX instances
• JavaScript – powers the site’s frontend logic
• MySQL – database used to store evidence
• NGINX – web server; one instance serves the main site while a separate CDN instance is used for images, videos, and audio files
• PHP – powers the site’s backend logic
• Traefik - reverse proxy providing HTTPS

Database schema

Since turnaround time was of utmost importance, I used the simplest method to store different communication types: each in their own table. This does not allow for easily adding new communication types, however that is a non-issue as no additional communication types have ever been needed.

Attachments are stored in a separate table to allow for a simple many-to-one relationship between texts / emails and attachments. The filename field stores the filename of the attachment in the filesystem which is accessible via the CDN.

Texts

Column Name	Description
id	unique ID
date_sent	date communication was sent
date_sent_year	generated column containing only year
date_sent_month	generated column containing only month number
date_sent_day	generated column containing only day number
from_address	phone number from which message was sent
to_address	phone number to which message was sent
victim	victim involved in communication (“name 1”, “name 2”, or “both”)
direction	direction of message (IN = sent to victim; OUT = sent from victim)
body	body of text
hash	MD5 hash of body/time/number used to detect potential duplicates

Emails

Column Name	Description
id	unique ID
date_sent	date communication was sent
date_sent_year	generated column containing only year
date_sent_month	generated column containing only month number
date_sent_day	generated column containing only day number
from_address	email address from which message was sent
to_address	email address to which message was sent
victim	victim involved in communication (“name 1”, “name 2”, or “both”)
direction	direction of message (IN = sent to victim; OUT = sent from victim)
subject	subject line from email
body	text-only body of email
html_body	HTML body of email (if applicable)
message_id	used in path to email attachments to prevent filename collisions
hash	MD5 hash of body/time/number used to detect potential duplicates during initial import

Voicemails

Column Name	Description
id	unique ID
date_sent	date communication was sent
date_sent_year	generated column containing only year
date_sent_month	generated column containing only month number
date_sent_day	generated column containing only day number
from_address	phone number from which message was sent
to_address	phone number to which message was sent
victim	victim involved in communication (“name 1”, “name 2”, or “both”)
direction	direction of message (IN = sent to victim; OUT = sent from victim)
body	transcription of voicemail
duration	duration of voicemail in seconds
filename	filename of voicemail for embedding in web interface
hash	MD5 hash of body/time/number used to detect potential duplicates during initial import

Videos

Column Name	Description
id	unique ID
date_sent	date video was posted
date_sent_year	generated column containing only year
date_sent_month	generated column containing only month number
date_sent_day	generated column containing only day number
from_address	YouTube account on which video was posted
victim	victim involved in communication (unused; always “both”)
direction	direction of message (IN = sent to victim; OUT = sent from victim)
title	title of video on YouTube
body	automatic transcription of video from YouTube
duration	duration of video in seconds
filename	filename of video for embedding in web interface

Attachments

Column Name	Description
attachment_id	unique ID / primary key
message_id	foreign key referencing either texts or emails tables
filename	filename of attachment

Web UI

A performant, user-friendly web interface is what made the project useful to law enforcement by allowing centralized searchable access to a multitude of communications across a multitude of platforms.

Impact

This system provided law enforcement with a structured, efficient way to review digital evidence, reducing time spent manually sorting through hundreds of scattered communications across many platforms. The alternative was hundreds and hundreds of screenshots and forwarded emails which quickly overloaded the investigator working the case.