HomeMy WebLinkAbout26 - Submissions - SS4A Demonstration Activity - Safety Data Platform (5)From:Nick Noreña
To:Bozeman Procurement
Cc:Sean Albert; Mark Buckner
Subject:[EXTERNAL][SENDER UNVERIFIED]SS4A Comprehensive Demonstration Activity - Safety Data Platform Proposalby Crow Flies, due Feb 19th, 3pm MT
Date:Thursday, February 19, 2026 1:14:04 PM
Attachments:SS4A Comprehensive Demonstration Activity Safety Data Platform Proposal.docx - 2_19_26, 11_57 AM.pdf
CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you
recognize the sender and know the content is safe.
For the Bozeman City Clerk's Office,
I would like to formally submit a proposal on behalf of the Crow Flies team for the Data
Safety Platform, a part of the SS4A Comprehensive Demonstration Activity. Please find ourPDF attached.
We'd greatly appreciate confirmation that you have received this proposal when you can
provide it, and please reach out directly if you have any questions. Thank you for yourconsideration.
Best,
-- Nick Noreña
Co-Founder & Product Lead @ Crow Fliescrowflies.dev
PROPOSAL
SS4A Comprehensive Demonstration Activity
Safety Data Platform
Submitted to
City of Bozeman, Montana
PO Box 1230, Bozeman, MT 59771
Submitted by
Crow Flies, LLC
https://crowflies.dev
February 19, 2026
Crow Flies, LLC - Safety Data Platform Proposal
1. Executive Summary
Crow Flies, LLC is pleased to submit this proposal to the City of Bozeman for the
implementation and three-year demonstration of a Safety Data Platform under the SS4A
Comprehensive Demonstration Activity. We propose to design, build, and support a custom
platform purpose-built for Bozeman’s specific needs: ingesting crash data from multiple sources,
generating visualizations and analysis tools, and directly supporting the development and
evaluation of the City’s Comprehensive Safety Action Plan, which is targeted for public
availability in December 2027.
We understand the City anticipates receiving proposals from commercially available off-the-shelf
platforms. We believe a purpose-built platform offers Bozeman a stronger outcome at this
budget. Commercial SaaS products are designed for broad markets and carry features,
complexity, and ongoing license costs that exceed what Bozeman needs. The City’s critical
requirements are specific and bounded: intake and centralize crash reports; extract and display
information from those reports across multiple formats; overlay crash data with roadway
characteristics and demographics; and analyze combinations of data to identify high-injury
networks and potential countermeasures.
Our platform will be built entirely on open-source foundations: a PostGIS geospatial database
for storing and querying crash data, open-source visualization and exploration tools such as
Apache Superset for dashboards and map-based analysis, automated data pipelines for
ingesting information from the City’s existing systems, and a crash risk scoring and
countermeasure analysis engine that processes spatial crash data to identify high-injury
networks and propose evidence-based safety improvements. This means the City of Bozeman
is not locked into a vendor contract beyond the three-year demonstration period. When the
contract ends, the City retains a platform built on industry-standard, open-source technology
that it can self-host, modify, or hand off to any developer, with no proprietary license fees and no
data trapped inside a vendor’s system. Based on the requirements laid out in the RFP and
the context we gathered from the questions we submitted, we have prototyped out what
this platform will be, and you can find a link to that prototype in Section 4 of our
proposal.
As a product studio specializing in public-sector and civic technology, Crow Flies brings deep
expertise in GIS development, data visualization, and user-centered design for developer tools.
Our team has direct experience building mapping and data platforms for government agencies
and public benefit organizations. We are proposing a collaborative, iterative development
process where City engineering staff will have access to working iterations of the platform
throughout the five-month development period, ensuring the platform reflects how Bozeman
staff actually work with crash data. Formal onboarding for senior stakeholders will occur at the
conclusion of development, once the platform is stable and feature-complete.
All implementation and support work will be completed within the $144,000 budget across the
three-year contract period.
Confidential
Page 2
Crow Flies, LLC - Safety Data Platform Proposal
2. Firm Profile
2.1 About Crow Flies
Crow Flies is a product studio that helps clients build the right software for the right people in the
right way. Founded in September 2025 with the three principals based in Bozeman, Tucson, and
Berkeley, Crow Flies partners with organizations that measure success not just in revenue, but
in social impact. We specialize in working with public-sector, civic-tech, and mission-driven
organizations, combining product strategy with quality technical execution.
Unlike typical development shops that execute on specs, or strategy consultants who disappear
after recommendations, Crow Flies bridges both worlds. We help organizations figure out what
to build and then build it right. Our team challenges assumptions, validates before, during, and
after building, and ensures every line of code serves actual user needs. Crow Flies offers
something a large SaaS vendor cannot: direct access to the people building the platform. During
the development period and beyond, City staff will work directly with the team that designed and
built the system.
Legal Name: Crow Flies, LLC
Type: Product Studio
Website: https://crowflies.dev
Team Size: 3 principals
Work Model: Distributed team
2.2 Relevant Service Capabilities
The following Crow Flies service capabilities are directly applicable to the Safety Data Platform:
• GIS Development: Deep expertise in Esri/ArcGIS, Mapbox, Leaflet, and PostGIS for
spatial databases. We design GIS solutions that serve specific use cases—whether
analyzing spatial patterns, communicating impact, or supporting data-driven decisions
about where to focus resources.
• Data Processing & Visualization: ETL pipeline development, data transformation, and
interactive dashboard creation using Python data science libraries, D3.js, and modern
visualization tools. We start by understanding what questions you need to answer, then
work backward to the data architecture.
• Full-Stack Development: End-to-end application development across Node.js, Python,
Go, React, Vue, PostgreSQL, and cloud platforms (AWS, GCP, Azure). We build
modern, scalable, and maintainable applications using industry best practices.
• API Integrations: Custom REST and GraphQL API development, enterprise platform
integration, and government/civic data source connectivity. We build integrations with
robust error handling, monitoring, and documentation.
2.3 Key Personnel
Sean Albert - Principal, Engineering Lead
Confidential
Page 3
Crow Flies, LLC - Safety Data Platform Proposal
Engineering leader with 15+ years building products for government and public benefit
organizations. Sean brings a proven track record scaling teams, shipping enterprise-grade
software, and implementing responsible AI frameworks for civic applications. On this project, he
will own the platform architecture, core development, and deployment infrastructure.
• Co-Founder & Engineering Lead at Crow Flies (2025–Present): Full-stack
development across the development lifecycle, including LLM and ML integrations, rapid
prototyping, and bringing new products to market.
• Lead Software Engineer at Exygy (2021–2024): Designed and implemented security
policies meeting FedRAMP-adjacent requirements for large city governments, including
the City of San José. Built features making it easier for over eight million residents
across multiple jurisdictions to find and apply for affordable housing. Collaborated with
Google.org fellowships on civic technology solutions.
• Lead Product Engineer at Simpleview Inc. (2018–2021): Led system design and
engineering management for the flagship CMS product. Built custom map building tools,
page builders, API integrations, and trained production developers.
Mark Buckner - Principal, Engineering Lead
Engineering leader based in Bozeman, MT with expertise across the full stack including
JavaScript/TypeScript, Python, SQL, React, and data pipeline engineering. Mark brings direct
experience building civic technology platforms and open-source government software. On this
project, he will focus on data pipeline engineering, frontend development, and visualization
implementation.
• Co-Founder & Engineering Lead at Crow Flies (2025–Present): Full-stack engineer
leading product development from concept through deployment, with experience across
data engineering, analytics infrastructure, and interactive data visualization. Building
scalable systems and intuitive tools that turn complex datasets into actionable insights.
• Senior Software Engineer II at Exygy (2022–2025): Developing the open-source
Bloom Affordable Housing System and collaborating with San Francisco Digital Services
on the DAHLIA platform. Led additional civic technology projects including a child care
grants platform for Alameda County using low-code tools, a data platform for the
Rosalynn Carter Institute, and a full-stack reentry support platform for formerly
incarcerated individuals.
• Senior Software Engineer at Triple Tree LLC (2018–2020): Full-stack engineering for
a technology consultancy using Node.js, React/React Native, and MySQL. Led a veteran
assistance platform project, managed client relationships, handled DevOps, and
integrated third-party tools including Salesforce and Airtable.
• Lead Instructor at Montana Code School (2017–2019): Taught full-stack web
development at a nonprofit coding school in Bozeman using a
React-Express-PostgreSQL stack. Delivered multi-month programs combining lectures
with public project demonstrations.
Nick Noreña - Principal, Product Lead
Confidential
Page 4
Crow Flies, LLC - Safety Data Platform Proposal
Product Manager with 12+ years developing products and educational content for
mission-driven organizations. Nick brings product positioning, stakeholder management, and
user research expertise. On this project, he will serve as the primary point of contact for the City
of Bozeman and lead all product strategy, requirements gathering, and stakeholder
coordination.
• Co-Founder & Product Lead at Crow Flies: Leads product strategy for clients looking
to ensure their products create observable impact and lead to key outcomes, as well as
product coaching and consulting services.
• Senior Product Manager at Exygy (2024–2025): Led product development for multiple
public-sector technology platforms serving underserved communities. Launched
Our415.org connecting vulnerable San Francisco families to critical resources.
Expanded Detroit’s affordable housing platform from ~185 to ~1,000 usable listings.
Managed the Bay Area’s Doorway Housing Portal, translating unclear client requests
into actionable roadmap items.
• Product Development & Strategy Coach (2023–2024): Worked with mission-driven
organizations including The Nature Conservancy, Nordic Innovation House, and Driver’s
Seat Cooperative. Led product management for Driver’s Seat, improving a key user
activation metric by 31%.
• Head of Coaching at Kromatic (2016–2022): As founding team member, created 40+
hours of educational content supporting a business model pivot that secured a multi-year
US Federal Government contract. Designed curriculum for 10 startup accelerators and
government innovation programs.
•
2.4 Team Roles & Responsibilities
Team Member Role Responsibilities
Sean Albert Platform Architect &
Engineer
Technical architecture, platform development,
data pipeline engineering, GIS implementation,
security, deployment, infrastructure
Mark Buckner Data & Visualization
Engineer
Data pipeline engineering, frontend development,
visualization implementation, PostGIS database
development, testing, documentation
Nick Noreña Product Manager & City
Liaison Primary City liaison, requirements gathering,
stakeholder coordination, user research, iterative
feedback sessions, acceptance testing
Confidential
Page 5
Crow Flies, LLC - Safety Data Platform Proposal
3. Scope of Services
3.1 Understanding of Need
The City of Bozeman adopted the Streets Are for Everyone (SAFE) Action Plan in 2022
following the tragic loss of two community members while cycling. Through SS4A Planning and
Demonstration Grants awarded in 2023 and 2024, the City is now strengthening this effort with a
Comprehensive Safety Action Plan (targeted for public availability in December 2027),
advanced safety data collection equipment at signalized intersections, and this safety data
platform.
Currently, City engineering staff must manually review individual crash reports to identify trends,
which is a labor-intensive process that limits the City’s ability to proactively utilize crash data.
The platform we propose will transform this workflow by centralizing crash data from multiple
sources, automating pattern analysis, and providing visualization tools that make safety insights
immediately accessible to staff.
We understand that the safety plan contract, the safety data platform, and the safety data
equipment contracts are all being awarded along the same timeframe, and that the safety plan
effort will take approximately 14 months. The platform must support the data analysis aspects of
the planning effort that are on the front end of the project. Our development approach is
designed to deliver a usable platform to City engineering staff as quickly as possible to
maximize its utility during plan development.
3.2 Why a Purpose-Built Platform
We believe a purpose-built approach offers Bozeman a stronger outcome for several reasons:
• Right-sized for Bozeman’s needs: Commercial SaaS platforms are designed for broad
markets and carry features, complexity, and ongoing license costs that exceed what
Bozeman needs. The City’s critical capabilities are specific: intake crash reports, extract
and display data from multiple formats including PDFs, overlay with roadway
characteristics and demographics, and identify high-injury networks. A purpose-built
platform delivers exactly this. Your budget goes directly toward features the City will
actually use.
• No vendor lock-in: Built entirely on open-source technology—a PostGIS geospatial
database and open-source visualization tools—the platform does not tie the City to a
proprietary vendor beyond the three-year demonstration period. This architecture also
allows the City to remain flexible with where they host the data, even opening up the
option of hosting it locally. PostGIS is the industry-standard geospatial database used
across government and enterprise GIS applications worldwide. When the contract ends,
the City retains a platform it can self-host, modify, or engage any developer to maintain,
with no escalating license fees and no data trapped in a vendor’s system.
• Timeline-aligned development: A commercial vendor requires the City to adapt to their
platform’s existing workflows. We build around the City’s timeline. Our iterative
development approach means engineering staff are using the platform for plan
development within months, not waiting for an enterprise onboarding cycle. We
coordinate directly with both City staff and the safety plan consultant team.
Confidential
Page 6
Crow Flies, LLC - Safety Data Platform Proposal
• Support right-sized to usage: Post-plan-development platform usage will be
concentrated around ongoing monitoring of plan metrics and evaluation of potential
projects during Capital Improvement Plan development. Our support structure is scaled
to this reality—responsive and available when needed, without enterprise-tier pricing for
a tool used periodically.
• Direct access to the builders: With Crow Flies, City staff work directly with the people
who designed and built the platform. There are no support tickets, no account managers,
no layers between staff and the engineering team.
3.3 Implementation Approach
We propose a collaborative, iterative development process structured across five months of
intensive development followed by 31 months of ongoing support. Our approach ensures City
staff are engaged throughout development, not just at the end.
Iterative Development with City Staff
City engineering staff involved in day-to-day crash data analysis will have access to working
iterations of the platform on a regular cadence throughout development. Starting as early as
Month 2, staff will be able to interact with progressive builds, test features against real
workflows, and provide feedback that directly shapes the next iteration.
This iterative approach serves three purposes: it ensures the platform reflects how Bozeman
staff and the other consultants supporting the Comprehensive Safety Action Plan actually work
with data; it surfaces usability issues early when they are inexpensive to fix; and it builds staff
familiarity with the platform organically so they are not encountering it for the first time at launch.
Formal onboarding for senior stakeholders and leadership will occur at the conclusion of the
development period, with the option of running an in-person training and onboarding session,
once the platform is stable and feature-complete. This ensures leadership sees a polished
product and receives training that reflects the final user experience. Ongoing onboarding for
new team members will be supported through documentation and training materials throughout
the support period.
Phase 1: Discovery & Architecture (Month 1)
• Kickoff meeting with City engineering staff and Comprehensive Safety Action Plan
consultant team
• Detailed requirements gathering and workflow mapping with staff who currently review
crash data
• Data source audit: obtain sample data from Zuercher Suite (now a part of Central
Square’s Public Safety Suite), AASHTOWare exports, and PDF crash reports
• Finalize technical architecture design, including data models, API design, PostGIS
database schema, ETL pipeline design, and infrastructure planning
• Establish development environment and CI/CD pipeline
Phase 2: Core Platform Development (Months 2–3)
• Data ingestion pipelines for Zuercher Suite, PDF parsing, and AASHTOWare formats
Confidential
Page 7
Crow Flies, LLC - Safety Data Platform Proposal
• Map-based crash visualization using Bozeman street network data from the City’s GIS
portal
• Crash diagram generation and basic pattern analysis
• Demographic data overlay using US Census Bureau data
• First iteration available to City engineering staff for hands-on testing and feedback
Phase 3: Advanced Features & Integration (Months 4–5)
• High-injury network identification and visualization
• Countermeasure recommendation engine
• SS4A grant evaluation reporting tools
• Integration with advanced safety data collection equipment (as equipment becomes
available from the separate procurement)
• Platform refinement based on ongoing staff feedback from iterative testing
• Senior stakeholder onboarding and formal training for all platform users
• User documentation and training materials
Phase 4: Ongoing Support & Enhancement (Months 6–36)
Following the five-month development period, we will provide ongoing platform support including
bug fixes, performance monitoring, security updates, implementation and operation support,
onboarding for new team members, and coordination with the Comprehensive Safety Action
Plan consultant team as their needs evolve.
The City has indicated it does not expect feature enhancement beyond the RFP requirements;
our support phase is structured accordingly, focused on reliable operation and resolving any
issues with required functionality. That said, we will be available to build out new ETL pipeline
components for data generated by the eventual advanced safety data collection equipment the
City is procuring.
Confidential
Page 8
Crow Flies, LLC - Safety Data Platform Proposal
4. Description of Proposed Solution
4.1 Platform Overview
The Bozeman Safety Data Platform is a web-based application purpose-built for City staff to
access, analyze, and act on crash data. The platform centralizes data from multiple sources into
a unified interface organized around interactive geospatial workspaces featuring multiple map
views, exploratory tools, and integrated analytics, with tools for pattern analysis, demographic
overlay, high-injury network identification, and countermeasure evaluation.
4.2 Open-Source Architecture
The platform is built on a four-layer open-source architecture designed specifically for the
geospatial analysis and visualization work this project requires. Rather than adopting an
off-the-shelf data catalog or portal tool, we are assembling purpose-built components that
directly address the City’s core needs: ingesting crash data, storing it in a format optimized for
spatial queries, and providing rich visualization and exploration capabilities for City staff and the
safety plan consultant team.
Layer 1: Data Ingestion (ETL Pipelines)
Automated pipelines will extract crash data from the City’s existing systems, transform it into
standardized formats, and load it into the geospatial database. These pipelines handle the data
quality, deduplication, and geocoding work that ensures the platform always reflects the most
complete and accurate crash data available. As new data sources come online, including the
advanced safety data collection equipment being procured through a separate RFP, new
pipeline components can be added without disrupting existing functionality.
How Layer 1 is built:
The platform will use Apache Airflow to orchestrate scheduled pipelines that pull crash records
from Zuercher Suite and AASHTOWare via API or direct database connections, parse PDF
reports using an OCR service like AWS Textract, and pass everything through dbt models that
handle standardization, deduplication, and geocoding before loading into the PostGIS database.
New data sources can be added simply by writing new Airflow DAGs and dbt models without
touching existing pipeline logic.
Layer 2: PostGIS Geospatial Database
All crash data, street network information, and demographic overlays will be stored in a PostGIS
database (PostgreSQL with geospatial extensions). PostGIS is the industry standard for
geospatial data management, used across government agencies, transportation departments,
and GIS applications worldwide. It is purpose-built for the complex spatial queries this project
likely demands, like identifying crash clusters at intersections, calculating corridor-level severity
scores, performing proximity analysis against road features, and overlaying demographic data
by geography.
PostGIS provides several critical advantages for this project:
Confidential
Page 9
Crow Flies, LLC - Safety Data Platform Proposal
• Optimized for geospatial analysis: Unlike general-purpose databases or data
warehouses, PostGIS natively supports spatial indexing, geometric operations, and
coordinate system transformations—the building blocks of crash pattern analysis and
high-injury network identification.
• Industry standard: PostGIS is compatible with virtually every GIS tool in use today,
including the City’s existing Esri/ArcGIS environment. Data stored in PostGIS can be
consumed directly by ArcGIS, QGIS, or any standards-compliant GIS application,
ensuring the platform complements rather than replaces the City’s current tooling.
• Open-source and low-cost: PostGIS is fully open-source with no licensing fees.
Hosting costs on a major cloud provider are minimal, and the City can self-host or
migrate the database to any environment at any time.
• No proprietary dependencies: Any developer with PostgreSQL experience can work
with this database. The City is never dependent on a single vendor for access to its own
data.
How Layer 2 is built:
The PostGIS database will be hosted on a managed PostgreSQL service which offers native
PostGIS support with minimal configuration. Tables will be designed around the analytical needs
of the platform - storing crash records, street network data, and any relevant geographic
overlays - with spatial indexes (GIST) applied to geometry columns to ensure fast query
performance as the dataset grows. Core PostGIS functions like ST_Buffer, ST_Intersects,
ST_DWithin, and ST_ClusterDBSCAN will power the platform's spatial workflows: cluster
detection, proximity analysis, corridor scoring, and High Injury Network identification. Because
the database speaks standard SQL and exposes data through well-documented schemas, it
can connect directly to the City's existing ArcGIS environment, eliminating the need for duplicate
data storage or manual exports.
Layer 3: Visualization and Exploration
The visualization layer provides City staff with the dashboards, interactive maps, and
exploratory analysis tools they need to turn raw crash data into actionable safety insights. We
will use open-source visualization and exploration tools such as Apache Superset to deliver:
• Interactive map visualizations: Crash locations plotted on Bozeman’s street network
using deck.gl geospatial visualization layers on top of the Mapbox platform. Among
others, key capabilities include spatial aggregation, generating heatmaps, and filtering
by severity, date range, crash type.
• Analytical dashboards: Pre-built views for temporal trends, spatial patterns, and
high-injury network summaries, with the ability to layer in demographic and street
network data from the PostGIS database.
• Ad-hoc data exploration: An explore mode that allows staff to group, filter, and
aggregate crash data by any dimension without requiring technical assistance - enabling
the kind of investigative analysis that supports plan development.
• Custom components: The ability to build bespoke features (such as countermeasure
scoring and recommendation tools) as custom widgets integrated directly into the
Confidential
Page 10
Crow Flies, LLC - Safety Data Platform Proposal
platform, tailored to the specific analytical workflows that emerge during plan
development.
• Self-hosted and City-owned: Deployed within the City’s cloud environment so all data
and all tooling remain under City control. No data leaves the City’s infrastructure.
How Layer 3 is built:
The visualization layer is built on Apache Superset, which provides the dashboards, interactive
maps, and ad-hoc exploration tools out of the box. Superset connects directly to the PostGIS
database and uses deck.gl with Mapbox for geospatial visualizations like heatmaps and filtered
crash maps. Where Superset's native capabilities fall short - such as countermeasure scoring
tools or custom analytical workflows - we build bespoke React components that embed directly
into the Superset interface, keeping everything within a single platform the City owns and
controls.
This architecture is deliberately flexible. Because all data flows through a standards-compliant
PostGIS database, the visualization layer can be supplemented or replaced over time without
affecting the underlying data. If the City's needs evolve, the foundation remains solid, and we
are able to react to those changes to implement the right features in the solution.
This open-source foundation also positions the platform to potentially serve broader data needs
within the recently designated Gallatin Valley Metropolitan Planning Organization (MPO). The
City's SS4A application notes that the safety data platform can be leveraged in the development
of the MPO's Long Range Transportation Plan and Unified Planning Work Program, extending
the impact of this investment across the wider MPO planning area.
Layer 4: Countermeasure Recommendations
The recommendations layer translates crash data and spatial analysis into actionable guidance
for City planners, engineers, and subject-matter experts. Rather than leaving staff to manually
interpret crash patterns, this layer surfaces specific, evidence-based countermeasure options,
such as signal timing changes, crosswalk upgrades, or speed reduction measures, all matched
to the conditions identified at a given location.
How Layer 4 is built:
This layer runs as a standalone Python microservice that queries the PostGIS database directly,
pulling crash records, severity scores, and spatial attributes for a given location or corridor.
Against that data, it applies a rules-based scoring engine that is weighted by factors like crash
frequency, severity index, road geometry, and pedestrian exposure, to rank applicable
countermeasures against an established framework such as FHWA's Proven Safety
Countermeasures. Scores and rankings are written back to the database, where they can be
consumed by the dashboard layer or exported for use in planning workflows.
Because the scoring logic is defined in code and operates on structured spatial data, it is both
auditable and extensible. This architecture also creates a clean integration point for AI: a
machine learning model could be substituted or layered into the scoring pipeline to refine
recommendations over time based on real-world outcomes, without requiring changes to the
surrounding system. Importantly, any AI-assisted recommendations would function as decision
support only; every suggestion would be reviewed and validated by City staff before any action
Confidential
Page 11
Crow Flies, LLC - Safety Data Platform Proposal
is taken. No countermeasure would be implemented without deliberate human sign-off, ensuring
that domain expertise and local judgment remain central to the process.
4.3 Data Ingestion
Data flows into the platform through automated ETL (Extract, Transform, Load) pipelines that
pull from the City’s existing systems, normalize the data into consistent formats, geocode crash
locations, and load everything into the PostGIS database. Each data source gets its own
pipeline component, making the system modular: when a new source comes online, a new
pipeline is added without disrupting the existing ones.
The City has indicated that a prioritized approach may be acceptable as long as the platform
utilizes the most complete crash data set available for plan development. We propose the
following initial priority order based on what is known today:
1. PDF Crash Reports (Priority 1): Structured data extraction from Montana crash report
PDFs using document parsing, extracting location, severity, contributing factors, vehicle
types, and other relevant fields. This is the most immediately available data source
based on sample reports provided by the City.
2. Zuercher Suite (Priority 2): Automated import of crash records from the City’s existing
records management system, coordinated with City IT staff during the discovery phase.
3. AASHTOWare (Priority 3): Import from Montana DOT’s crash data system, supporting
standard AASHTOWare export formats to ensure the most complete crash data picture.
4. Advanced Safety Data Collection Equipment (Priority 4): Integration with intersection
safety monitoring equipment as it is procured and deployed by the City through the
separate equipment RFP. We are capable of ingesting data in a variety of
industry-standard formats (JSON, XML, CSV, API endpoints) and will coordinate with the
equipment vendor once selected to determine the appropriate integration approach.
Importantly, we do not expect to know every data source or format on day one. Working with a
small, responsive team like Crow Flies means we have the flexibility to adapt as requirements
emerge during the discovery phase and throughout plan development. If the safety plan
consultant team identifies a new data source that would strengthen the analysis, or if the City’s
data landscape shifts as the equipment vendor comes online, we can build a new pipeline
component and integrate it without rearchitecting the platform. This adaptability is a core
advantage of the modular, open-source approach over a rigid commercial product where
integrations are constrained by the vendor’s roadmap.
4.4 Core Features
The following features address the critical capabilities identified by the City: intake and
centralize crash reports; extract and display information from those reports; overlay crash data
with roadway characteristics and demographics; and analyze combinations of data to identify
high-injury networks, high crash locations, and locations that present as potential safety
concerns.
Confidential
Page 12
Crow Flies, LLC - Safety Data Platform Proposal
Interactive Crash Map: Map-based visualization of crash incidents on Bozeman’s street
network, with filtering by date range, severity, crash type, weather conditions, and other
attributes. Powered by the City’s existing GIS data from public-bozeman.opendata.arcgis.com.
Crash Diagrams: Automated generation of crash diagrams from ingested data showing
collision dynamics, vehicle movements, and contributing factors at specific locations.
Pattern Analysis: Statistical analysis of crash patterns including temporal trends (time of day,
day of week, seasonal), spatial clustering, contributing factor correlation, and before/after
comparisons for interventions.
Demographic & Street Network Overlays: Overlay crash data with US Census demographic
data (including household income, vehicle access, and population density) and street
classification information to identify equity considerations and infrastructure correlations in crash
patterns.
High-Injury Network Identification: Automated identification and visualization of corridors and
intersections with disproportionately high crash rates, weighted by severity, to prioritize safety
investments.
Countermeasure Recommendations: Evidence-based countermeasure suggestions linked to
identified crash patterns, drawing from FHWA Proven Safety Countermeasures and relevant
research.
SS4A Grant Evaluation Support: Built-in reporting tools to produce evaluation metrics and
documentation required for the SS4A grant quarterly performance progress reports, tracking
progress against safety targets.
4.5 Safety Data Platform Prototype
As a part of our submission, we have developed a prototype of the platform we are proposing.
The prototype is built on the same PostGIS + Apache Superset open-source technology stack
that the full platform will use, connected to real geospatial data drawn from Bozeman's street
network and populated with representative crash data.
The sections below walk through what the prototype includes and how each component maps to
the platform capabilities described in this proposal.
We recommend following this guide when exploring the prototype. Here is the link to the
live prototype - https://bzn-safety-data-portal-production.up.railway.app/superset/dashboard/1/.
The login credentials for this prototype are:
● (u) admin
● (p) admin
Confidential
Page 13
Crow Flies, LLC - Safety Data Platform Proposal
The Dashboard: A Single Unified View
The prototype is organized as a single dashboard titled “Bozeman Safety Data Platform” which
reflects our intent to give the City staff one place to go for all crash data analysis. While the
prototype has limited data, we wanted to populate it with a view that a City staff member might
see when opening the dashboard at their desk.
Dashboard Detail: KPI Summary Tiles
Across the top of the dashboard, three summary tiles give staff an immediate situational picture
without requiring any navigation or analysis. This prototype includes KPI Summary Tiles derived
from sample charts and metrics created in the platform:
• Total Crashes (31): The total number of crash records in the current filtered view. This
number updates live when any filter is applied — a staff member can immediately see how
many crashes occurred in a given year, season, or intersection area.
• # of HIN Intersections (10): The count of intersections currently designated as part of the
High-Injury Network — those with disproportionately high crash rates weighted by severity.
This is a key metric for the safety plan, giving leadership an immediate answer to the
question: how many locations require priority attention?
• % in Dark Lighting Conditions (17): The share of crashes that occurred under dark or
poor lighting conditions — a metric that directly informs lighting improvement
countermeasures. This is the kind of targeted, actionable insight that previously required
manual cross-referencing of individual crash reports.
These tiles demonstrate a core design principle of the platform: the most important numbers are
always visible, always current, and require no analysis by the person reading them.
Temporal Pattern Analysis: Crashes by Hour of Day
Adjacent to the KPI tiles, a bar chart displays crash frequency by hour of day. This chart makes
immediately visible the temporal patterns that safety engineers need to understand: when are
crashes happening, and does the pattern point toward enforcement windows, signal timing
adjustments, or lighting improvements?
In the full platform, this analysis will extend to day-of-week breakdowns and seasonal trends,
giving the safety plan consultant team the temporal data they need to design targeted
countermeasures and evaluate the before-and-after impact of any interventions the City
implements.
Geospatial Crash Visualization: Interactive Map
The crash map below the tiles and bar chart plots crash density across Bozeman's street
network. Hexagonal cells aggregate nearby crashes into a spatial summary: taller, more
intensely colored hexagons indicate higher crash concentrations. This format is particularly
effective for identifying crash clusters at specific intersections and corridors, and creates the
visual output that can drive High-Injury Network identification.
Confidential
Page 14
Crow Flies, LLC - Safety Data Platform Proposal
The map is interactive: City staff can zoom into specific corridors, hover over hexagons to see
underlying crash counts, and apply date range or other filters to see how crash density shifts
over time or under different conditions. This is the platform replacing hours of manual report
cross-referencing with a visual, immediately navigable spatial picture.
The underlying dataset, v_crashes_with_road, joins crash records with Bozeman's street
network data, meaning every crash point is spatially associated with the road segment it
occurred on. This is what enables the platform to calculate corridor-level metrics and severity
scores rather than treating crashes as unconnected point events.
Evidence-Based Recommendations: Countermeasure Analysis Table
Below the crash map, the Countermeasure Table is the most analytically sophisticated
component of the prototype. It demonstrates a capability that no off-the-shelf crash data
platform offers out of the box: an automated countermeasure recommendation engine that links
identified crash patterns at specific intersections to evidence-based safety interventions. With
this being a prototype, please note that the City staff will be able to guide the recommendations
engine to ensure the recommendations align with the City’s intent and goals, and also that the
recommendations surface novel ideas that might aid in solving a safety problem.
Each row in the table represents a specific countermeasure recommended for a specific
intersection, and includes the following attributes (see column headers):
• Intersection and HIN Tier: Which location, and whether it falls in the High or Low tier of
the High-Injury Network.
• Estimated Crash Reduction: The projected percentage reduction in crashes if the
countermeasure is implemented, drawn from FHWA Proven Safety Countermeasures
research.
• Category and Cost Tier: Whether the intervention is an Enforcement, Engineering, or
Operations measure, and its relative cost band, giving the City an immediate sense of
budget implications alongside effectiveness.
• Fatal and Crash Counts: The severity data that justifies the recommendation, surfaced
directly in the table.
• Status: Whether the countermeasure is Proposed or flagged for Priority Review, which
supports the City's project prioritization workflow.
• Rationale: A plain-language explanation of why this specific countermeasure was
recommended for this location, grounded in the crash pattern data. For example, the
recommendation of a Leading Pedestrian Interval at Durston Road and 19th Avenue is
explained by the high concentration of fatal crashes with glare and dusk conditions as
primary factors.
This table is an example of what the safety plan consultant team will use to develop the City's
countermeasure portfolio. Rather than producing this analysis manually from raw crash report
data, the platform generates it automatically from the PostGIS database, and updates it as new
crash data is ingested.
Confidential
Page 15
Crow Flies, LLC - Safety Data Platform Proposal
High-Injury Network Visualization: HIN Map
At the bottom of the dashboard, a scatter map uses a deck.gl Scatterplot layer to mark specific
intersections by their High Injury Network designation. Red dots indicate High-tier HIN
intersections; blue dots indicate Low-tier intersections. Both are plotted against Bozeman's
street grid on a light Mapbox basemap, making it easy to see the geographic distribution of
safety-critical locations across the city.
This map is powered by the v_hin_scatter dataset, a PostGIS database view that
pre-computes which intersections qualify for HIN designation based on severity-weighted crash
scores. The HIN tier designation drives the Countermeasure Table above: only High and Low
HIN intersections appear in the recommendation engine. This creates a coherent analytical
pipeline entirely within the platform: crash data flows into the PostGIS database, the HIN view
identifies priority locations, and the countermeasure view generates recommendations for those
locations.
The Data Architecture Behind the Dashboard
The five datasets powering the prototype dashboard can be found in the Datasets panel of the
Superset interface, and they illustrate how the platform's PostGIS database is structured to
support different analytical views of the same underlying crash data:
Dataset What It Powers
v_crashes_with_road The primary crash dataset, joining raw crash records with Bozeman road
network data. Powers the crash map, total crash KPI, and
crashes-by-hour chart.
v_high_injury_network A severity-weighted view that calculates HIN designation for each
intersection. Powers the # of HIN Intersections KPI.
v_hin_scatter A point dataset of HIN intersections with tier classification, optimized for
the scatter map visualization.
v_countermeasure_dash
board
The countermeasure recommendation view, linking HIN intersections to
FHWA-grounded interventions with estimated costs and crash reduction
projections.
pedestrian_crashes A filtered view of crashes involving pedestrians, powering the % in Dark
Lighting Conditions KPI.
Each dataset is a PostgreSQL view that lives inside the PostGIS database rather than a static
file or spreadsheet. When new crash data is loaded into the database (whether from PDF
parsing, Zuercher Suite, or AASHTOWare), every view updates automatically, and every chart
and metric on the dashboard reflects the new data the next time it is loaded. This is the core
operational advantage of the architecture: there is no manual data preparation step for City staff.
Scope of the Prototype
Confidential
Page 16
Crow Flies, LLC - Safety Data Platform Proposal
The prototype is a demonstration of platform capability, not a complete production system. For
the purposes of this submission, it intentionally focuses on the visualization and analysis layer,
which are the components most directly relevant to City staff's day-to-day work. It uses
representative data rather than live connections to City systems. Specifically, the prototype does
not yet include:
• Live ETL pipelines: Automated ingestion from Zuercher Suite, AASHTOWare, or PDF
crash reports. These will be built during Phase 1 and 2 of development using Apache
Airflow and dbt.
• Authentication and access control: The production platform will include role-based
access control and single sign-on integration with the City's existing credentials.
• Full crash record volume: The prototype is populated with representative data to
demonstrate analytical workflows. The production system will ingest the City's complete
historical crash record set during the discovery and data migration phase.
• Countermeasure Engine: The prototype has a static, manually generated
countermeasure table. In the production environment, this will be automatically generated
by an algorithmic or AI-assisted workflow, or whatever technical design comes out of
discussing more detailed requirements with City staff.
The linked prototype maps to the architecture diagram below, which shows how the full platform
connects the ETL ingestion layer, PostGIS database, and Superset visualization layer into a
complete system.
Crash Data Platform Architecture - A three-layer system that ingests crash data from multiple
sources (Zuercher Suite, AASHTOWare, PDFs, and city datasets) through Airflow-orchestrated
ETL pipelines, stores and spatially enriches it in a PostGIS database, and delivers insights via
Apache Superset dashboards and interactive maps. A dedicated Countermeasure Engine
Confidential
Page 17
Crow Flies, LLC - Safety Data Platform Proposal
interfaces with the database to score crash records and recommend safety interventions. Note:
This is the complete architecture to the proposed solution. The prototype only implements the
PostGIS database and visualization layer.
4.6 Security
The platform is built entirely on open-source technologies, giving the City full transparency into
the software stack and complete autonomy in how the system is deployed, configured, and
secured. Because the system can be self-hosted or deployed within a reputable cloud
environment, the City retains control over infrastructure-level security decisions, including
network configuration, access controls, and data residency.
The architecture is designed to support environments that require SOC 2 auditability. While the
platform itself is open-source software and not a certified product, it can be deployed within a
SOC 2–compliant hosting environment and configured to align with the Trust Services Criteria,
including:
● Access Controls – Role-based access control (RBAC), least-privilege permissions, and
authentication integration (SSO/OIDC/SAML as required).
● Audit Logging – Comprehensive logging of user access, data modifications, and
administrative actions.
● Encryption – Encryption in transit (TLS 1.2+) and encryption at rest through database
and infrastructure configuration.
● Data Integrity Controls – Controlled ETL pipelines with validation, versioning, and
reproducibility of analytical outputs.
● Change Management – Version-controlled codebase, documented deployment
processes, and traceable configuration updates.
● Availability & Resilience – Backup strategies, infrastructure redundancy, and disaster
recovery procedures defined at the hosting layer.
Because the system is open-source and modular, it can be deployed within an environment that
undergoes independent SOC 2 audits without architectural modification. This ensures the
platform meets enterprise security expectations while maintaining flexibility and long-term
sustainability for the City.
5. Maintenance and Support for City Staff
5.1 Support During Development (Months 1–5)
During the five-month development period, City engineering staff will have direct, ongoing
access to the Crow Flies team. You will have a dedicated team member serving as the primary
point of contact who will conduct regular check-in meetings with City staff. Working iterations of
Confidential
Page 18
Crow Flies, LLC - Safety Data Platform Proposal
the platform will be available for staff testing starting in Month 2, with structured feedback
sessions incorporated into the development cycle.
5.2 Ongoing Support (Months 6–36)
The City has indicated it does not expect feature enhancement beyond the RFP requirements,
and that platform usage outside of safety plan development will be limited to a few hours per
month for ongoing monitoring of plan metrics and evaluation of potential projects during Capital
Improvement Plan development. Our support structure is designed around this reality:
• Bug fixes and issue resolution
• Platform monitoring and performance optimization
• Security updates and patching
• Implementation and operation support, including resolving issues with required
functionality
• Onboarding support for new City staff members
• Coordination with the Comprehensive Safety Action Plan consultant team
• Integration support for advanced safety data collection equipment as it is deployed
• Any data integration work that comes from the procured safety data collection equipment
Support is likely to be highest during the initial safety plan development period when platform
usage is most intensive. As usage patterns stabilize, support effort will naturally decrease, and
our pricing reflects this trajectory rather than locking the City into enterprise-tier support costs for
periodic use.
5.3 Service Level Expectations
Crow Flies commits to 99.5% uptime for the Safety Data Platform, measured monthly, excluding
scheduled maintenance windows. Scheduled maintenance will be performed during
non-business hours with advance notice. Response times for requests will range from same day
for critical issues to two business days for non-critical issues. Business hours are 9am-5pm MT.
5.4 Onboarding & Training
Our onboarding approach is designed around two tracks:
Ongoing Staff Engagement (Months 2–5): City engineering staff who work directly with crash
data will have access to iterative platform builds throughout development. This hands-on
engagement builds familiarity and ownership of the platform organically. These team members
will provide regular feedback that shapes platform development, ensuring the final product
reflects their real workflows.
Senior Stakeholder Onboarding (Month 5): At the conclusion of development, we will conduct
formal onboarding sessions for senior stakeholders and leadership. This ensures they
experience the platform in its final, polished form and receive training that reflects the complete
feature set. We will produce comprehensive user documentation and training materials at this
stage.
Confidential
Page 19
Crow Flies, LLC - Safety Data Platform Proposal
New Staff Onboarding (Months 6–36): Throughout the support period, we will provide
onboarding assistance for any new City staff members who need access to and training on the
platform. User documentation and training materials will be maintained and updated as the
platform evolves.
6. Related Experience with Projects Similar to the Scope of
Services
6.1 Crow Flies: SnowObs / National Avalanche Center
The Snowbound team, creators of the SnowObs avalanche forecasting platform, reached a
critical technical bottleneck as they scaled to meet the needs of institutional users like the
Alaska DOT. Their platform was already saving forecasters over 30 minutes daily, but their
internal team was stretched thin by support demands and organizational transitions. They
needed a partner who could jump into their four-microservice environment and immediately
unblock high-priority roadmap items that had been stalled for months.
Our approach is to build with the Snowbound team as a persistent extension of their
engineering muscle. This partnership began with a successful two-month engagement in late
2025, where we integrated a Field Work Plan (FWP) feature for the National Avalanche Center's
Avalanche Forecast Platform (AFP), used by 23 avalanche centers and serving millions of
public users each year. The FWP allows forecasters to generate and share plans including
map-based location data, multi-role user management, one-click check-ins, robust filtering, and
shareable password-protected views.
That initial work established a foundation of high trust and technical alignment, proving that
Crow Flies could navigate Snowbound's existing stack (PostgreSQL with PostGIS, Python, Vue)
and deliver production-ready code while maintaining their long-term product vision.
Since then, we've remained integrated directly into Snowbound's workflows, moving seamlessly
between strategic planning and execution. Two key deliverables are now complete:
● AvalancheObs Drawing: We built and shipped the frontend map-drawing tools wired to
existing backend geometry storage, eliminating the friction of manual KML uploads and
allowing forecasters to record observations natively on the map.
● Token Authentication Architecture: We designed and deployed a scalable
token-based API access system, providing a secure, maintainable foundation for
Snowbound to share data with institutional partners without accruing technical debt.
By unblocking these roadmap delays, we've positioned Snowbound to secure enterprise
revenue from clients requiring sophisticated data integrations. As we move through early 2026,
our focus shifts to completing the production rollout of remaining mapping features and
establishing long-term API infrastructure. This ongoing collaboration allows the Snowbound
team to stay focused on their mission of avalanche safety while Crow Flies ensures the
technical foundation is robust, secure, and ready for the next phase of institutional growth.
This project is directly relevant to the Safety Data Platform because it involved building
map-based visualization tools for a government-adjacent safety organization, working with
Confidential
Page 20
Crow Flies, LLC - Safety Data Platform Proposal
geographic data to support life-safety decisions, and designing interfaces for field professionals
who need to quickly access and act on spatial data.
6.2 Team Experience: Rosalynn Carter Institute
Designed and built a multi-platform data ecosystem spanning Google BigQuery, Looker, and a
purpose-built platform to administer caregiving programs at scale. At the core of the platform
was a custom ETL pipeline that extracted programmatic data from operational systems and
external sources (including CDC public health datasets) transformed and standardized it across
schemas, and loaded it into BigQuery as a centralized analytical data warehouse. This
aggregation layer was essential to making disparate data sources comparable and queryable at
scale, enabling the analytics engines built on top of it to assess participant outcomes and
program efficacy with consistency and rigor. Custom visualization tools in Looker surfaced those
insights for program creators and administrators in an interpretable, real-time view, without
requiring technical expertise to operate. By automating data flows that had previously been
manual and fragmented, the platform meaningfully reduced administrative overhead while
enabling the kind of evidence-based evaluation that funders and program leaders increasingly
require.
6.3 Team Experience: Exygy - Civic Technology Platforms
Exygy is a public benefit corporation that builds technology to advance equity in housing,
government services, and civic infrastructure. Their portfolio spans multiple high-impact
platforms serving millions of residents, and their work frequently involves complex data
pipelines, geolocation services, and integrations with government systems.
Crow Flies principals served as core engineering and product team members at Exygy across
several of these platforms over a four-year period.
On Our415.org and HousingReadinessReport.org, principals provided software engineering and
product management on data-intensive platforms that leveraged geolocation services to
connect residents with local resources. Both projects required building and managing complex
data pipelines to ensure accurate, location-aware information delivery.
On Bloom Housing, an open-source affordable housing platform, principals served as technical
lead and product manager. Bloom streamlines the affordable housing application process for
both applicants and housing agencies, and is now used across multiple jurisdictions.
On CiviForm, the open-source benefits platform originally developed by Google.org and
recognized as one of Time's Best Inventions of 2025, a principal served as the initial technical
lead. In that role, they implemented a GIS-based eligibility feature that determines whether
applicants qualify for programs based on geographic service areas. CiviForm now helps over 8
million residents access government benefits, and the platform's data architecture supports
complex eligibility logic across multiple jurisdictions and program types.
This body of work demonstrates deep experience with the data engineering, GIS integration,
and multi-stakeholder platform challenges that are central to building a safety data platform.
Crow Flies principals bring firsthand knowledge of designing systems where location data,
eligibility logic, and government requirements intersect.
Confidential
Page 21
Crow Flies, LLC - Safety Data Platform Proposal
7. References
The following contacts have worked with Crow Flies team members on the projects listed above:
Scott Havens, Managing Director at Snowbound Solutions LLC
scott@snowboundsolutions.com
Ke Wang, Data Analyst at Rosalynn Carter Center
ke.wang@cartercenter.org
Zach Berke, CEO of Exygy
zach@exygy.com
8. Proposed Schedule
The following schedule reflects Crow Flies’ anticipated timeline, assuming a contract start date
of April 2026. The schedule is designed to ensure the platform is available for the front-end data
analysis aspects of the Comprehensive Safety Action Plan, which the City anticipates will take
approximately 14 months with the final plan document publicly available in December 2027.
Timeline Phase Key Deliverables
Month 1 Discovery & Architecture Requirements doc, architecture design,
development environment, data source
audit, PostGIS database and ETL
pipeline setup
Months 2–3 Core Platform Development Data ingestion pipelines, crash map,
pattern analysis, demographic overlays.
First iteration to engineering staff for
testing.
Months 4–5 Advanced Features & Launch High-injury networks, countermeasures,
SS4A reporting, senior stakeholder
onboarding, user documentation
Months 6–36 Ongoing Support Bug fixes, monitoring, security updates,
operation support, new staff
onboarding, equipment integration as
available
Confidential
Page 22
Crow Flies, LLC - Safety Data Platform Proposal
9. Price Proposal
9.1 Total Price
Crow Flies proposes a total fixed price of just under $144,000 for the full scope of services
described in this proposal, covering the 36-month contract period.
9.2 Price Breakdown
Phase Duration Monthly Rate Subtotal
Development & Implementation 5 months $15,000 $75,000
Ongoing Support & Maintenance 31 months $2,225 $68,975
Total 36 months $143,975
9.3 What’s Included
The fixed price includes all costs associated with delivering the scope of services, including:
platform development, cloud hosting and infrastructure for the full 36-month period, all software
tools and libraries (no third-party license fees since the platform is built on open-source
technology), staff training and documentation, ongoing support and maintenance, and
coordination with the safety plan consultant team and equipment vendor.
This pricing reflects one of the key advantages of the open-source approach: there are no
recurring software license fees embedded in the support cost. The entire platform—PostGIS
database, visualization tools, ETL pipelines—is built on open-source technology. The monthly
support rate covers Crow Flies team time for maintenance, monitoring, and issue resolution.
9.4 Payment Schedule
Crow Flies proposes monthly invoicing at the rates outlined above, with payment due within 30
days of invoice. Invoices will be issued on the first business day of each month for the prior
month's services.
Confidential
Page 23
Crow Flies, LLC - Safety Data Platform Proposal
10. Affirmation of Nondiscrimination
Crow Flies, LLC hereby affirms it will not discriminate on the basis of race, color, religion, creed,
sex, age, marital status, national origin, or because of actual or perceived sexual orientation,
sexual preference, gender identity, or disability in fulfillment of this contract. This prohibition on
discrimination shall apply to the hiring and treatment of Crow Flies’ employees and to any and
all subcontracts entered into in the fulfillment of the services identified herein.
In addition, Crow Flies, LLC hereby affirms it will abide by the Equal Pay Act of 1963 and
Section 39-3-104, MCA (the Montana Equal Pay Act).
Name and title of person authorized to sign on behalf of submitter:
Sean Albert, Co-Founder
Crow Flies, LLC
Date:
Confidential
Page 24
Crow Flies, LLC - Safety Data Platform Proposal
11. Proposed Changes to the Software as Service Agreement
11.1 Items Accepted as Written
Crow Flies accepts the Software as Service Agreement substantially as presented, including
Section 9 (Indemnity/Waiver of Claims/Insurance), which the City has indicated is not subject to
modification. We also note that Section 18 (Intellectual Property Ownership) clearly delineates
that the Provider retains all IP rights to the platform while the City retains full ownership of City
Data. We accept these terms as written.
11.2 Proposed Modifications
• Section 10 — Audit (SOC 2 Type II): This section requires an annual SOC 2 Type II
report. Obtaining a SOC 2 Type II audit is prohibitively expensive for a small firm and
would consume a disproportionate share of the project budget. We propose an
alternative: the platform will be built exclusively using infrastructure and tools that are
themselves SOC 2 Type II compliant (e.g., AWS, managed database services,
authentication providers). This ensures City data is handled within SOC 2–compliant
environments without requiring the Provider to independently obtain the certification. We
are happy to provide documentation of the compliance certifications held by each
infrastructure and tool provider used in the platform.
Confidential
Page 25
Crow Flies, LLC - Safety Data Platform Proposal
12. City of Bozeman Cloud Services Questionnaire
1. Service Levels
Crow Flies commits to 99.5% uptime for the Safety Data Platform, measured monthly, excluding
scheduled maintenance windows. Scheduled maintenance will be performed during
non-business hours with advance notice to City staff.
2. Data Ownership
Per Section 18 of the Software as Service Agreement, the City of Bozeman retains full
ownership of all City Data submitted, posted, or otherwise transmitted through the platform.
Crow Flies retains ownership of the platform IP (Provider IP), including the software,
documentation, and aggregated statistics. Crow Flies will not access or use City Data for any
purpose other than as described in the Agreement. The City’s data will be stored in a PostGIS
database using standard, open formats, ensuring the City can extract and use its data
independently at any time using any PostgreSQL-compatible tool.
3. ADA Compliance
The platform will be developed to meet WCAG 2.1 Level AA accessibility standards. This
includes proper semantic HTML, keyboard navigation, screen reader compatibility, color
contrast ratios, and alternative text for visual elements. Accessibility will be tested during
development and prior to launch.
4. Data Security
All data is encrypted in transit (TLS 1.2+) and at rest (AES-256). The platform will be hosted in a
single-tenant environment, meaning City data is not co-located on shared hardware with other
customers. No PII beyond what is contained in crash reports will be gathered by the platform.
No credit card transactions will occur through the platform; PCI compliance is not applicable.
5. Data Integrity
Data integrity is maintained through automated database backups (daily), transaction logging,
input validation on all data ingestion pipelines, and version tracking. The ETL pipeline
architecture ensures all incoming data passes through validation and transformation steps
before reaching the PostGIS database, catching formatting errors, duplicates, and anomalies at
the point of ingestion. All data modifications are logged with timestamps and user attribution.
6. Data Location
All City data will be stored exclusively in data centers located within the United States. Crow
Flies will not transfer City data outside the United States without prior written consent from the
City.
7. Responding to Legal Demands
In the event Crow Flies receives a subpoena or legal request for the City’s data, we will promptly
notify the City in writing or by phone before disclosing any data, unless legally prohibited from
doing so. Crow Flies will cooperate with the City’s legal counsel to respond appropriately and
will not release City data to third parties without the City’s authorization unless compelled by
court order.
Confidential
Page 26
Crow Flies, LLC - Safety Data Platform Proposal
8. Data Breach Reporting
In the event of a Data Incident, Crow Flies will notify the City in writing or by phone within 48
hours of discovery, consistent with the Agreement’s Data Incident provisions. The notification
will include a description of the incident, the types of data affected, the steps taken to contain
the breach, and remediation measures. Crow Flies will cooperate fully with the City in
investigating and responding to any breach.
9. Disaster Recovery
The platform infrastructure includes automated daily backups stored in a geographically
separate US-based facility, point-in-time recovery capability, and documented failover
procedures. Recovery Time Objective (RTO) is 24 hours. Recovery Point Objective (RPO) is 24
hours (one business day of data). Backup restoration is tested periodically to ensure reliability.
10. Business Continuity / Exit
In the event of contract termination or business discontinuation, Crow Flies will provide the City
with a complete export of all City Data in standard, open formats (CSV, GeoJSON, Shapefile,
and PostgreSQL database dump). Because the platform is built entirely on open-source
technology (PostGIS, open-source visualization tools, Python-based ETL pipelines), the City is
not dependent on Crow Flies for continued operation. PostGIS is the industry-standard
geospatial database—any developer or GIS professional with PostgreSQL experience can take
over platform hosting, maintenance, and modification. Crow Flies will provide 90 days of
transition support, including knowledge transfer, documentation, and assistance with migration
to the City’s chosen hosting environment or successor provider.
This is a fundamental advantage of building on industry-standard open-source technology: the
City’s exit strategy is built into the platform architecture, not dependent on a vendor’s willingness
to cooperate.
11. Termination Policy
Termination provisions are governed by Section 3 of the Software as Service Agreement. Either
party may terminate for material breach with a 60-day cure period. The Provider must notify the
City 90 days in advance of the Agreement’s expiration date. No expiration or termination affects
the City’s obligation to pay fees that became due before termination, nor entitles the City to
refunds for services already rendered.
Questionnaire Completed by: Date:
Confidential
Page 27