Understand communities. Support better decisions.

TrendSights

Community-level research.
In Less time.

TrendSights is an AI-assisted research workspace for understanding communities. Explore public data, compare places through supported workflows, and create maps, notebooks and briefs with sources and methods attached.

Explore research outputs Request beta access

For community researchers and analysts supporting planning and program decisions.

Severe City Street crashes in Seattle, Bellevue and Tukwila, aggregated into one-kilometer cells for 2021–2023. Darker cells contain more reported crashes. — WSDOT records · 2021–2023 · Three-city study in King County. Counts are not adjusted for travel exposure.
Inspect the map and sources

Define a question and place

Start with the decision you need to support. Set the geography, period and measures that make the question answerable.

Explore supported public data

Work with available sources and workflows. Keep definitions, coverage and uncertainty visible as you explore.

Review and share results

Inspect the findings and limits, then use maps, notebooks and briefs to help colleagues follow the evidence.

A recorded introduction · TOP Sprint 2026

Meet TrendSights in 2:18

Rob Schaper · Tech Team Lead, TOP Sprint 2026

See how a research question connects to public data, maps and portable research files, with sources, methods and human review kept visible.

Prepared for U.S. Census Open Innovation Labs, this recorded development demonstration follows a saved community study and the evidence behind its outputs.

Recorded October 19, 2026 · 2:18 · English captions included. Available beta workflows may differ from this development demonstration.

Watch on YouTube (opens in a new tab)

Meet TrendSights

My name is Rob Schaper. I lead the CloudRaven tech team here with CloudRaven Labs, and I'm happy to present TrendSights, a CloudRaven Labs community research program delivered as part of the 2026 U.S. Census Open Innovation Labs TOP sprint focused on federal AI data readiness.

Scope the research

Here we are starting with reusable verified TrendSights workflow for community level research on a target topic. We can see the question targeted, source inputs, proposed artifacts and research workflow, school planning, economic change, transportation risk. Communities all kind of grapple with similar needs, but struggle with unique constraints and demographics of their community.

Saved study and counts

So opening a saved Kitsap County transportation study, for example, we can quickly see that we've got a number of fatal crashes and people killed across three years. But where are the hotspots? The annual table from the workflow keeps those counts and population context distinct. So downstream experiments can be further developed in production systems like Databricks and Snowflake.

Place and uncertainty

Researchers want to move into different levels of detail as well. So moving from county map to census tract, they also want to look at the individual records. So mapping counts and the results and the community estimates, we want to keep them all together, which we do in this example. So we want to include things like margins of error so that statistical significance testing can take place.

Sources and records handoff

Researching the source data takes on the form of a ledger fused to a workflow method. So federal data, periods, the geography, the review state, it all stays visible. It all stays observable, which is really helpful in building a workflow for something like a public records request.

Portable research artifacts

So again, the theme here with TrendSights is observability and portability of the artifacts that we generate as researchers. We can take the evidence with us, geometry only, PDF, CSV tables, data dictionaries, for example. We have the given workflow, but this all gets built into a notebook, which we can download and use internally for offline validation or in other systems.

Human review and workspace

So human review within managed agentic research, these workflows drives a new format of community level research. And that's TrendSights, a managed community research platform by CloudRaven Labs. If you'd like to request a workspace today and get started, please go to trendsights.io. Thank you.

From analysis to an answer others can inspect

Explore the research outputs TrendSights supports.

Make community evidence easier to understand, check and reuse. Explore maps, notebooks and briefs, then see how reviewed work can be presented in a publication. The examples come from CloudRaven’s King County study.

MapsSee the place and context

Research BriefsRead findings and limits

Evidence AtlasData Science Notebooks + visual evidence

Embedded ArtifactsBrowse a reviewed publication

Maps · Place and context

See how the evidence varies by place.

Maps help you compare places and understand the geography behind a finding. In this King County example, resident and workplace maps add context to the crash question across Seattle, Bellevue and Tukwila.

Explore the map study

Published maps comparing resident density and workplace primary job density in Seattle, Bellevue and Tukwila. — Actual study artifact · ACS resident context and LODES workplace jobs

Locally developed CloudRaven publications and a reproducibility notebook from the published study tables illustrate these output types and the publication experience shown here. Available outputs depend on the supported workflow; the complete study was developed separately from hosted TrendSights runs.

A worked research example · King County, Washington

What do roads, residents and workplaces add to a crash study?

Community research often needs more than one source to explain a pattern. CloudRaven combined crash records, Census resident data and workplace context across Seattle, Bellevue and Tukwila to illustrate the kinds of research outputs TrendSights supports.

Start with the rise in severe city-street crashes, then inspect the maps, source definitions and published findings. Change the city, road category and time detail to see what the evidence can support.

Explore the case study Read the two-page brief

Crashes: 2020–2024 · Increase focus: 2021–2023
ACS: 2020–2024 · Primary jobs: 2023 · Updated October 20, 2026

King County locator and a map of Seattle, Bellevue and Tukwila with severe City Street crash counts aggregated into one-kilometer cells, 2021–2023. — Locally developed CloudRaven research · Federal, state and local evidence connected through a documented workflow. Crash counts are not adjusted for travel exposure.

A method you can follow

Keep the question, data and findings connected.

TrendSights organizes research around CRISP-DM, a six-stage process for moving from a question to reviewed results. See what each stage contributes, illustrated by CloudRaven’s King County study.

01 · Business Understanding

Define what you need to know.

Define the community question, places, period and intended use before choosing data or methods.

King County example: severe city-street crashes in Seattle, Bellevue and Tukwila, focused on 2021–2023.

What you carry forward: A clear question and a bounded research scope.

Published King County study map locating Seattle, Bellevue and Tukwila and the severe City Street crashes included in the research scope. — King County study · Three-city scope, 2021–2023View full size

Revisit earlier stages as evidence changes. The locally developed King County artifacts illustrate the CRISP-DM process; a researcher reviews findings before use.

Another way to begin · Private plugin beta

Already use Codex? Start your question there.

TrendSights Research skills offer Codex users an additional entry path: scope a community question, identify sources and choose a supported research workflow.

Use the TrendSights workspace to inspect the available results and artifacts. Review sources, methods and limitations before sharing the work with colleagues.

01Scope the question in Codex.
02Choose an available research workflow.
03Review and share the resulting work.

Request plugin beta access

Beta access is reviewed. Ask about plugin availability and supported workflows for your setup.

TrendSights Research plugin configuration header showing the plugin identity and private, unpublished status. — Reproducibility example from published study tables. Download notebook (.ipynb)

City	2021	2023	Change
Seattle	179	259	+80
Bellevue	20	21	+1
Tukwila	12	14	+2

Your next research question

Bring your community question to TrendSights.

Request a private beta workspace for your research. Create and verify your account, share your research needs, and have your access request reviewed. Account creation and workspace approval are separate steps.

Request beta access Watch the walkthrough

Research perspective · 4 min read

How CRISP-DM Could Help Community Researchers Manage Insight Fatigue

Agentic research can generate findings faster than people can review them. CRISP-DM could help community researchers keep the question, evidence and next action connected.

Read the article